Content uploaded by Julia Riley
Author content
All content in this area was uploaded by Julia Riley on Dec 28, 2019
Content may be subject to copyright.
© 2019 The Linnean Society of London, Biological Journal of the Linnean Society, 2019, XX, 1–14 1
Biological Journal of the Linnean Society, 2019, XX, 1–14. With 5 gures.
Uncovering the function of an enigmatic display:
antipredator behaviour in the iconic Australian frillneck
lizard
CHRISTIAN A. PEREZ-MARTINEZ1,*,†,, JULIA L. RILEY2,‡, and MARTIN J. WHITING1,
1Department of Biological Sciences, Macquarie University, Sydney, NSW 2109, Australia
2Ecology and Evolution Research Centre, School of Biological, Earth and Environmental Sciences,
University of New South Wales, Sydney, NSW 2052, Australia
†Current address: Division of Biological Sciences, University of Missouri, Columbia, MO 65211, USA
‡Current address: Department of Botany and Zoology, Stellenbosch University, Stellenbosch, Western
Cape 7600, South Africa
Received 26 August 2019; revised 29 October 2019; accepted for publication 31 October 2019
When faced with a predator, some animals engage in a deimatic display to startle the predator momentarily,
resulting in a pause or retreat, thereby increasing their chance of escape. Frillneck lizards (Chlamydosaurus kingii)
are characterised by a large, pronounced frill that extends from the base of the head to beyond the neck and, when
displayed, can be up to six times the width of the head. We used behavioural assays with a model avian predator to
demonstrate that their display conforms to deimatic display theory. First, juveniles and adults deployed the frill in
encounters with a model predator. Second, the display revealed three colour patches (white and red–orange patches
on the frill; yellow mouth palate) that facilitate a transition from a cryptic to a conspicuous state as perceived by a
raptor visual system. Third, the display was performed with movements that amplified its effect. The frill area was
larger in males than in females, which suggests that the frill might also be co-opted for male–male contests. If future
research confirms a role of the frill in male agonistic interactions, frillneck lizards will be a rare case in which a
structure has a dual function in a deimatic display and a sexually selected signal.
ADDITIONAL KEYWORDS: Chlamydosaurus kingii – conspicuous display – crypsis – deimatic display – reptile –
sensory ecology – squamate – startle display – visual display – visual modelling.
INTRODUCTION
The mechanisms involved in antipredator strategies
are diverse; animals use crypsis and masquerade to
circumvent detection or recognition, aposematism
and mimicry to broadcast unpalatability, and motion
dazzle and flicker fusion to obscure movement
trajectories (Stevens & Merilaita, 2009; Umeton
et al., 2017; Loeffer-Henry et al., 2018). These
adaptations are all tailored for perception by the
visual system of a predator. The leaf-like masquerade
of a phasmid, the myrmecomorphy in appearance and
locomotion of a jumping spider and the conspicuous
wing pattern of a monarch butterfly all function in
portraying a certain morphological state to predators
(MacDougall & Dawkins, 1998; Wedmann, 2010;
Nelson, 2012). Furthermore, some animals display
different phenotypes depending on the identity of
the observer. For example, dwarf chameleons exhibit
different colour changes for camouflage in response
to bird and snake predators (Stuart-Fox et al., 2006,
2008). However, most studies have focused on static
signals, such as aposematism, whereby signals are
broadcast continuously (Mappes et al., 2005; Stevens &
Ruxton, 2012; but see Rowe & Halpin, 2013), whereas
signalling coupled with a behavioural response yields
a multifaceted strategy during a confrontation with
a predator. In the late stages of a predatory threat,
some animals will employ a marked transition from
an inconspicuous to a highly conspicuous state,
called a deimatic display (Maldonado, 1970). This
unprecedented transition aims to elicit a ‘startle’
applyparastyle “g//caption/p[1]” parastyle “FigCapt”
*Corresponding author. E-mail: perez.christian.alessandro@
gmail.com
Downloaded from https://academic.oup.com/biolinnean/advance-article-abstract/doi/10.1093/biolinnean/blz176/5679583 by University of Stellenbosch user on 19 December 2019
2 PEREZ-MARTINEZ ET AL.
© 2019 e Linnean Society of London, Biological Journal of the Linnean Society, 2019, XX, 1–14
response in a predator, resulting in a pause or retreat,
thereby increasing the performer’s chance of escape
(Umbers et al., 2015; Umbers & Mappes, 2016).
Deimatic displays are likely to be more common than
we think, but they are often overlooked in comparison
to other antipredator behaviour and colouration, such
as aposematism and camouflage (Umbers et al., 2017).
In contrast to warning signals, deimatic displays are
dynamic, because they involve a rapid transition from
a cryptic to a conspicuous state and do not require
predator learning or innate aversion (Umbers et al.,
2017; Holmes et al., 2018). Therefore, species that use
deimatic displays appear cryptic when in the normal
active or resting state because they may, for example,
show strong background matching. This behavioural
transition from a resting to a display state can be
dramatic, and the effect may be amplified by behaviour
(e.g. inflation of the body, hissing, gaping of the mouth).
The end result is an intimidating display aimed to
induce a reflexive response in the receiver (predator).
Frillneck lizards (Chlamydosaurus kingii Gray, 1825)
are unique among squamates because they have a large,
extensible frill that substantially increases the apparent
size of the lizards when erect, with the exception of small
head flaps in the toad-headed agama (Phrynocephalus
mystaceus Pallas, 1776; MJW, unpublished observations).
Frillneck lizards have long been known to extend their
frills outwards (Saville-Kent, 1896); however, despite the
novelty of the frill, its function has remained enigmatic.
Early hypotheses of the behavioural context of the frill
included gliding, food storage, auditory enhancement,
crypsis and thermoregulation, none of which have received
any empirical or observational support (De Vis, 1883;
Fenner, 1933; Bacchus, 1939; Worrell, 1963; Frith & Frith,
1987; Shine, 1990). Only two hypotheses of the function
of the frill may be of significance: its use in antipredator
behaviour and as a social signal (Shine, 1990). Current
behavioural data are insufficient to test the hypothesis
of the role of the frill as an antipredator mechanism.
Previous data describe a lack of sexual dimorphism in
the frill (Shine, 1990), which casts doubt on its role in
intrasexual selection.
Here, we hypothesised that the defensive display
of the Australian frillneck lizard (Chlamydosaurus
kingii) conforms to the predictions of deimatic display
theory. We predicted that: (1) the frill is erected in
response to the presentation of a model predator;
(2) exposure of colour patches on the frill creates a
rapid transition from a cryptic to a conspicuous state
according to the visual system of a predator; and (3) the
display is performed in conjunction with behaviours
that amplify its effect. We also measured the size of the
frill and conspicuousness through the visual system of
a lizard to examine differences in sexual dimorphism
that might suggest a role in signalling to conspecifics.
MATERIALS AND METHODS
Study SyStem
The frillneck lizard is a large, diurnal agamid [mean
snout–vent length (SVL) 25.4 cm in males and 20.7 cm
in females; Shine, 1990] that is locally abundant
in savannah woodland habitats across northern
Australia, extending from West Australia to southern
Queensland and including southern Papua New
Guinea (Shine, 1990; Griffiths & Christian, 1996;
Cogger, 2002). They are arboreal sit-and-wait predators
that prefer habitats with dense canopy cover, a low
density of shrubs, and grassy vegetation (Griffiths &
Christian, 1996). The most notable feature of frillneck
lizards is the extensible frill, in our data spanning up
to six times the width of the head and supported by
hyoid cartilage.
Fieldwork was conducted from October 2017 to
March 2018 at Fogg Dam Conservation Reserve
(hereafter, Fogg Dam; 12°35′40.2″S, 131°16′25.1″E)
in tropical northern Australia. All methods used were
conducted under Northern Territory research permit
no. 61517 and approved by Macquarie University
Animal Ethics Committee (reference no. 2017/046).
Fogg Dam is located along the Adelaide River floodplain
and experiences a distinct seasonal wet–dry climate.
The study ran from the early to late wet season, with
total monthly rainfall varying from 90 to 743 mm and
mean monthly daytime temperatures ranging from
28.8 to 32.3 °C (Bureau of Meteorology, 2018). Frillneck
lizards show higher activity and growth rates during
the wet season, when insect prey is plentiful and
lizards are reproductively active (Christian & Bedford,
1995; Ujvari et al., 2015).
Lizards were captured by hand or by noose at
night while they slept vertically on eucalypt trunks.
For each individual, we measured head height (the
distance between the parietal eye and the lower jaw),
head width (the distance between the two sides of the
lower jaw at the widest part of the head) and head
length (the distance between the quadrate–articular
jaw joint on the right side to the tip of the snout) to
the nearest 0.01 mm using digital callipers, in addition
to SVL and tail length to the nearest 1 mm using a
standard ruler. We calculated frill area by taking 25
measurements of the left half of the frill, summing
the area of the 12 triangular areas, and doubling
the result (Supporting Information, Fig. S1). Lizards
were sexed by head dimensions (see Supporting
Information, Supplementary methods), and any lizard
with an SVL < 180 mm was considered to be a juvenile
(corresponding to ~1 year of age; Christian et al., 1999).
Capture locations were recorded on a GPS (Garmin
GPSMAP 64s), and all lizards were released at their
point of capture within 24 h.
Downloaded from https://academic.oup.com/biolinnean/advance-article-abstract/doi/10.1093/biolinnean/blz176/5679583 by University of Stellenbosch user on 19 December 2019
FRILLNECK LIZARD ANTIPREDATOR BEHAVIOUR 3
© 2019 e Linnean Society of London, Biological Journal of the Linnean Society, 2019, XX, 1–14
Behavioural aSSayS and Scoring
We conducted behavioural assays on 52 lizards (14
males, 22 females and 16 juveniles) at Fogg Dam.
Individuals were placed in a cubic arena with wall
lengths of 1.5 m. The arena was positioned under a
large mimosoid tree with pockets of sunlight, providing
sufficient illumination while ensuring that the lizards
did not overheat. The sides of the arena were opaque
and the top uncovered, meaning that the lizard only
had a view of the canopy. All assays were conducted
in sunny conditions from 09.00 to 13.00 h. We
constructed a model raptor by printing and laminating
a digital image of the ventral surface of a peregrine
falcon (Falco peregrinus Tunstall, 1771) in flight and
attaching the centre of the cut-out to a 3 m extendable
pole. When shaken, the body of the raptor (29 cm long)
remained in position with the stick, while the wings
(71 cm wingspan) flapped back and forth with ~80° of
movement per wing. Down-welling light eliminated
glare on the underside of the model and illuminated
its outline and pattern.
Each lizard was presented with two treatments
separated by 2–3 h, in a random order that was
balanced: a model raptor as described above, and a
control consisting of a pole without the raptor model
attached. Before each assay, the lizard was placed on
the ground of the arena in a cloth bag, the bag was
lifted, and the lizard was given an acclimation period
of 15 min. Either the raptor model or the pole was
then placed 0.5 m above the arena (2 m above the
ground), and the pole was repetitively moved up and
down at a rate of about two flaps per second. The assay
occurred for 30 s, and each lizard was subsequently
placed back in the cloth bag after this period. Two
cameras (Sony FDR-X3000) were mounted on opposite
top corners of the enclosure and live-streamed to a
handheld iPhone 5s through the PlayMemories Mobile
application. This allowed us to orient the model raptor
or the pole directly above the lizard for the duration
of the assay. The lizard and experimenter were not
within view of one another.
From the footage, we scored whether the frill was
deployed (details below) in the control and experimental
treatments. In the bird model trials, we also scored
the number of lunges (L), the number of times the
lizard ran/fled from the model (R), and the duration
(in seconds) that the lizard had its frill fully erect (F),
frill partly erect (P), mouth agape (M) and remained
stationary (S). Other associated behaviours included
hissing and tail whipping, but these behaviours were
difficult to dissociate from other body movements
and noises; therefore, they were not quantified in the
behavioural scoring. All video footage was scored in
BORIS (Friard & Gamba 2016), as was the matrix
of transitions after each behaviour. We combined all
behaviours scored to summarize the display behaviour
of each individual lizard using eqn:
Behavioural score =
ï
(
1
/
2
)(
F
+
M
)+(
P
+
2L
)
30 ò−ïS
+
2R
30 ò
Mouth agape and frill fully erect are only performed
in concert with one another; therefore, we halved the
summed duration of the two. We considered variables
F, M, P and L to be defensive behaviours, whereas
S and R were more associated with a ‘flight’ or non-
defensive antipredator response. Lizards with a more
vigorous display would have a greater display score
than lizards with a fleeing or unreactive response.
To assess our summary of display behaviour, we also
summarized the behaviours, excluding flee, using
a principal components analysis (PCA), and the
indices agreed strongly (see Supporting Information,
Supplementary methods).
Frillneck diSplay colouration
‘Colour’ is an interaction of the light environment in
the habitat, the spectral reflectance of the animal and
the visual system of the receiver (Kemp et al., 2015;
Endler & Mappes, 2017), the last of which can only
be approximated without behavioural verification.
Herein, we use the term for convenience.
We used an Oceanoptics Jaz reflectance
spectrophotometer with an illumination probe
connected to a PX-2 light source to take spectral
reflectance measurements of eight body regions of 52
frillneck lizards. Spectra were standardised by taking
measurements at an angle of 90° and 5 mm from the
surface, covering an area of 6 mm2. Measurements
were relative to a dark and a 99% white (WS-1)
reflectance standard (Labsphere, Inc.). All raw spectra
were obtained from wavelengths 300–700 nm, which
encompasses the visual system of lizards and birds
(Fleishman et al., 2011; Lind et al., 2013) and were
averaged over 5 nm intervals using a kernel smoothing
function. We used the application OceanView to
retrieve the spectral data, which was processed in R
v.3.3.1 using the package PAVO (Maia et al., 2013).
We measured the spectral reflectance of the
following regions: dorsum (two locations); edge of the
frill (three locations); and the interior of the front of
the frill (three locations; Supporting Information, Fig.
S2). For all areas, three measurements were averaged
from the same colour patch. The regions along the edge
of the frill are visible when the frill is folded (i.e. when
the lizard is not in a display state) and are adjacent to
the interior colour patches when the frill is extended
outwards. The interior of the frill is characterised by
a red–orange patch on the lower part of the frill and
a white circular patch on the upper part of the frill.
Studies have examined similar spectrophotometric
Downloaded from https://academic.oup.com/biolinnean/advance-article-abstract/doi/10.1093/biolinnean/blz176/5679583 by University of Stellenbosch user on 19 December 2019
4 PEREZ-MARTINEZ ET AL.
© 2019 e Linnean Society of London, Biological Journal of the Linnean Society, 2019, XX, 1–14
measures of frillneck lizards (Merkling et al., 2016;
McLean et al., 2019), although, to the best of our
knowledge, palate spectra have not been measured
previously. The opening of the frill and mouth are
behaviourally coupled, and neither was observed
in isolation; therefore, the palate is a potentially
important component of the frillneck display.
To quantify the conspicuousness of colour patches,
we modelled spectra through both avian and lizard
visual perspectives (see ‘Colour analysis’ section
below). For each lizard, we used an average of spectra
from multiple body regions to represent the resting and
display states. The colour of a lizard in a resting state
was quantified using two regions along the dorsum
and three along the edge of the frill. As arboreal
agamids, frillneck lizards occupy vertical eucalypt
trunks when in a resting state (Griffiths & Christian,
1996; Supporting Information, Fig. S2), and avian
predators are most likely to spot them from perches.
When examined from an avian perspective, from
above, the white patches are positioned laterally and
are not visible; therefore, we did not include the white
patches in the resting state (Supporting Information,
Fig. S2). The measures of conspicuousness for the
resting state were taken as the average of two regions
on the dorsum and three regions on the edge of the
frill against the average reflectance of 30 samples
of eucalypt bark. Bark samples were collected from
the study site at perches from which the lizards had
previously been captured.
The colour of a lizard when in the display state
was quantified using the red–orange patch, the white
patch and the palate, which were analysed separately
(Supporting Information, Fig. S2). The red–orange
patch is almost entirely confined to the bottom three
frill folds and is obscured when the lizard is at rest.
Two elliptical white patches are positioned directly
to the sides of the open mouth and exposed palate.
Measures of conspicuousness for the display state
were taken as the three interior colour patches against
the average of the three adjacent regions on the edge
of the frill.
viSual modelling
We modelled frillneck lizard colouration through
the visual system of a bird predator. Whistling kites
(Haliastur sphenurus Vieillot, 1818) and black kites
(Milvus migrans Boddaert, 1783) are abundant birds of
prey at Fogg Dam (Sergo & Shine, 2015) and are likely
to be the main predators of frillneck lizards. Given
that microspectrophotometry data are unavailable
for these species, we chose the common buzzard
(Buteo buteo Linnaeus, 1758) as a close relative in the
Accipitriformes to model raptor vision.
To explore the hypothesis that the frill plays a role
in male–male signalling whereby male frills will be
more conspicuous than in females, we also modelled
colouration as perceived by a lizard conspecific. We
used photoreceptor sensitivity data from the ornate
crevice-dragon (Ctenophorus ornatus Gray, 1845),
which is also an agamid, because lizard visual systems
are conserved (Fleishman et al., 2011) and no data
exist currently for the frillneck lizard.
To estimate the ability of an avian predator or lizard
conspecific to discriminate between different colour
patches, we applied Vorobyev–Osorio receptor-noise
models (Vorobyev & Osorio, 1998). The model uses
signal intensity (photoreceptor quantum catches) and
receptor noise to estimate distances in perceptual space
between spectra in units of ‘just-noticeable differences’
(JNDs; see Supporting Information, Supplementary
methods for model specifications and details).
StatiStical analySeS
All statistical analyses were performed in R v.3.3.1
(R Core Team 2018). Before all analyses, we explored
the data following the protocol of Zuur et al. (2010) for
validity of test assumptions and integrity of the data.
During this process, missing values were removed
from the data set. Most notably, palate spectra were
not recorded for one individual, and this explains the
discrepancy in sample size and degrees of freedom
between the models (see Tables 1 and 2). Also
before analyses, we used a rank-transformation to
normalise our behavioural display score (Riley et al.,
2017).
First, we wanted to examine whether the lizards
(males, females and juveniles combined) erected the
frill in response to a predator. To accomplish this, we
used Pearson’s χ
2 test (chisq.test function in the stats
R package; R Core Team, 2018) with Yates’ continuity
correction (Stefanescu et al., 2005) to compare
presence/absence of frill erection in 51 lizards between
the experimental and control assays.
Second, we examined differences between males,
females and juveniles in conspicuousness (in JNDs) of
the white patch and red patch of the frill, the palate and
the dorsum for raptor (B. buteo) and lizard (C. ornatus)
visual systems, in addition to frill area (in square
centimetres) and display behavioural scores. For each
of these response variables, we ran linear models,
using the function lm in the R package stats (R Core
Team, 2018), that included the continuous variable of
SVL and the categorical variable of sex [three levels:
female, male and juvenile (indeterminate sex)]. Models
were run initially with an interaction effect between
SVL and sex, but if this effect was not significant the
models were re-run without the interaction effect to
allow accurate interpretation of the main effects. If
Downloaded from https://academic.oup.com/biolinnean/advance-article-abstract/doi/10.1093/biolinnean/blz176/5679583 by University of Stellenbosch user on 19 December 2019
FRILLNECK LIZARD ANTIPREDATOR BEHAVIOUR 5
© 2019 e Linnean Society of London, Biological Journal of the Linnean Society, 2019, XX, 1–14
Table 1. Least-squares means and pairwise comparisons of chromatic and achromatic contrasts of colour patches according to the visual system of the common
buzzard (Buteo buteo), representing an avian predator
Least-squares means Pairwise differences of contrast
Sex Mean SE d.f. 95% Confidence interval Contrast βSE d.f. t P-value
Lower bound Upper bound
Chromatic contrast
White patch Female 4.260 0.575 47 3.103 5.416 Female–juvenile −7.674 1.707 47 4.495 < 0.001
Juvenile 11.934 1.519 47 8.877 14.990 Female–male 4.369 1.497 47 2.918 0.005
Male −0.109 1.475 47 3.076 2.857 Juvenile–male 12.043 2.826 47 4.261 < 0.001
Red patch Female 7.095 0.786 47 5.514 8.676 Female–juvenile −1.011 2.333 47 −0.433 0.667
Juvenile 8.106 2.076 47 3.929 12.283 Female–male 4.127 2.047 47 2.017 0.050
Male 2.968 2.015 47 −1.087 7.023 Juvenile–male 5.138 3.863 47 1.330 0.190
Palate Female 5.806 0.548 46 4.702 6.910 Female–juvenile −5.085 1.710 46 −2.974 0.005
Juvenile 10.891 1.570 46 7.731 14.051 Female–male 4.021 1.541 46 2.609 0.012
Male 1.785 1.491 46 −1.216 4.787 Juvenile–male 9.106 2.904 46 3.135 0.003
Dorsum Female 2.416 0.440 47 1.531 3.301 Female–juvenile −6.942 1.306 47 −5.315 < 0.001
Juvenile 9.359 1.162 47 7.020 11.697 Female–male 4.680 1.146 47 4.085 < 0.001
Male −2.264 1.128 47 −4.534 0.006 Juvenile–male 11.623 2.163 47 5.374 < 0.001
Achromatic contrast
White patch Female 11.958 0.556 47 10.839 13.078 Female–juvenile −6.512 1.652 47 −3.942 < 0.001
Juvenile 18.470 1.470 47 15.513 21.427 Female–male 3.893 1.449 47 2.687 0.010
Male 8.066 1.427 47 5.195 10.936 Juvenile–male 10.405 2.735 47 3.805 < 0.001
Red patch Female 6.363 0.581 47 5.193 7.532 Female–juvenile −5.760 1.726 47 −3.337 0.002
Juvenile 12.123 1.536 47 9.033 15.213 Female–male 2.910 1.514 47 1.922 0.061
Male 3.453 1.491 47 0.453 6.452 Juvenile–male 8.670 2.858 47 3.034 0.004
Palate Female 12.075 0.655 46 10.756 13.394 Female–juvenile −4.198 2.044 46 −2.054 0.046
Juvenile 16.273 1.876 46 12.496 20.050 Female–male 2.151 1.842 46 1.168 0.249
Male 9.924 1.782 46 6.337 13.511 Juvenile–male 6.349 3.471 46 1.829 0.074
Dorsum Female 4.131 0.617 47 2.891 5.372 Female–juvenile −3.105 1.830 47 −1.696 0.097
Juvenile 7.236 1.629 47 3.959 10.513 Female–male 2.194 1.606 47 1.367 0.178
Male 1.937 1.581 47 −1.244 5.118 Juvenile–male 5.299 3.030 47 1.749 0.087
Bold P-values indicate significance at a level of α = 0.05.
Downloaded from https://academic.oup.com/biolinnean/advance-article-abstract/doi/10.1093/biolinnean/blz176/5679583 by University of Stellenbosch user on 19 December 2019
6 PEREZ-MARTINEZ ET AL.
© 2019 e Linnean Society of London, Biological Journal of the Linnean Society, 2019, XX, 1–14
Table 2. Least-squares means and pairwise comparisons of chromatic and achromatic contrasts of colour patches according to the visual system of the ornate
crevice-dragon (Ctenophorus ornatus), representing a frillneck lizard conspecific
Least-squares means Pairwise differences of contrast
Sex Mean SE d.f. 95% Confidence interval Contrast βSE d.f. t P-value
Lower bound Upper bound
Chromatic contrast
White patch Female 4.883 0.853 47 3.167 6.599 Female–juvenile −7.638 2.532 47 −3.016 0.004
Juvenile 12.520 2.253 47 7.987 17.054 Female–male 4.443 2.221 47 2.000 0.051
Male 0.440 2.187 47 −3.961 4.840 Juvenile–male 12.081 4.192 47 2.882 0.006
Red patch Female 6.220 0.816 47 4.578 7.863 Female–juvenile −1.841 2.424 47 −0.760 0.451
Juvenile 8.061 2.157 47 3.722 12.401 Female–male 3.948 2.126 47 1.857 0.070
Male 2.272 2.094 47 −1.940 6.484 Juvenile–male 5.789 4.013 47 1.443 0.156
Palate Female 8.500 0.853 46 6.782 10.217 Female–juvenile −5.947 2.661 46 −2.235 0.030
Juvenile 14.447 2.443 46 9.529 19.365 Female–male 5.242 2.398 46 2.186 0.034
Male 3.257 2.320 46 −1.413 7.928 Juvenile–male 11.190 4.519 46 2.476 0.017
Dorsum Female 2.963 0.898 47 1.156 4.770 Female–juvenile −8.288 2.667 47 −3.108 0.003
Juvenile 11.251 2.373 47 6.477 16.025 Female–male 4.221 2.339 47 1.804 0.078
Male −1.258 2.304 47 −5.892 3.377 Juvenile–male 12.509 4.415 47 2.833 0.007
Achromatic contrast
White patch Female 17.863 0.808 47 16.238 19.489 Female–juvenile −6.098 2.399 47 −2.542 0.014
Juvenile 23.961 2.135 47 19.667 28.256 Female–male 4.141 2.104 47 1.968 0.055
Male 13.722 2.072 47 9.553 17.891 Juvenile–male 10.239 3.972 47 2.578 0.013
Red patch Female 11.127 0.851 47 9.416 12.838 Female–juvenile −5.337 2.525 47 −2.114 0.040
Juvenile 16.464 2.247 47 11.943 20.985 Female–male 3.313 2.215 47 1.496 0.141
Male 7.814 2.181 47 3.425 12.202 Juvenile–male 8.651 4.181 47 2.069 0.044
Palate Female 17.907 1.079 46 15.736 20.078 Female–juvenile −3.124 3.364 46 −0.929 0.358
Juvenile 21.030 3.088 46 14.814 27.247 Female–male 1.041 3.032 46 0.343 0.733
Male 16.866 2.933 46 10.962 22.770 Juvenile–male 4.165 5.713 46 0.729 0.470
Dorsum Female 4.874 0.843 47 3.178 6.570 Female–juvenile −8.696 2.503 47 −3.475 0.001
Juvenile 13.570 2.227 47 9.090 18.050 Female–male 6.167 2.195 47 2.809 0.007
Male −1.293 2.162 47 −5.642 3.056 Juvenile–male 14.863 4.143 47 3.587 < 0.001
Bold P-values indicate significance at a level of α = 0.05.
Downloaded from https://academic.oup.com/biolinnean/advance-article-abstract/doi/10.1093/biolinnean/blz176/5679583 by University of Stellenbosch user on 19 December 2019
FRILLNECK LIZARD ANTIPREDATOR BEHAVIOUR 7
© 2019 e Linnean Society of London, Biological Journal of the Linnean Society, 2019, XX, 1–14
a main effect of sex was found to be significant, we
used a post hoc test of least-squares means (using
the lsmeans R package with the functions lsmeans for
main effect comparisons and lstrends for interaction
effect comparisons; Lenth, 2016) to examine pairwise
differences of all contrasts between the sexes. For
all models, before interpretation, we verified the
assumptions of normality and homoscedasticity of
residuals. All linear models were set to a significance
level of α = 0.05.
RESULTS
Frill diSplay and Behaviour
Lizards erected their frills in response to the model
avian predator more often than in response to the
control pole (χ
2 = 34.16, d.f. = 1, P < 0.001). Of the 52
lizards, 43 (83%) of them remained unresponsive to
the control pole. The other eight lizards (17%) erected
their frills with their mouths agape in response to the
control pole. In response to the avian model, 38 lizards
(73.1%) erected the frill with their mouth agape, three
lizards (5.8%) fled, and the remaining 11 lizards (21.2%)
remained unresponsive. All lizards that erected the
frill initially did so < 0.5 s after presentation of the
model. Of the 38 lizards that deployed the frill initially,
50% lunged, 13.2% fled and 18.4% stopped displaying
(i.e. remained stationary). Of the 19 lizards that
lunged, 63.2% subsequently fled. No behaviours were
considered after ‘flee’ in the behavioural sequence
because the boundary of the arena might have affected
their natural behaviour (Fig. 1).
colouration From the perSpective oF a raptor
The red patch, white patch and palate of lizards (Fig.
2) were all distinguishable from their respective
contrast regions from the visual perspective of a
raptor; JNDs were greater than a discrimination
threshold value of one (Fig. 3). In addition, the
brown dorsum was distinguishable against the tree
background.
Just-noticeable differences of chromatic contrast
were higher for females than males in the white
patch, red patch, palate and dorsum. Juvenile JNDs of
chromatic contrasts were higher than for females and
males in the white patch, palate and dorsum (Table
1; Fig. 4). Just-noticeable differences of achromatic
contrast were higher in females than in males in only
the white patch. Juvenile JNDs of achromatic contrast
were higher than for females in the white patch, red
patch and palate, and higher than for males in both
the white patch and the red patch (Table 1). All other
colour patches were not different between sexes.
Chromatic and achromatic contrast in the white patch
and the red patch, in addition to dorsum chromatic
contrast, were significantly related to SVL (Supporting
Information, Table S1). No other colour patches were
related to SVL (Supporting Information, Table S1).
colouration From the perSpective oF a lizard
Through an agamid visual system, JNDs were higher
for females than for males in palate chromatic contrast
and dorsum achromatic contrast (Table 2). Juveniles
were greater than both males and females in JNDs
of chromatic contrast in the white patch, palate and
dorsum, and in achromatic contrast in the white
patch and the red patch (Table 2). All other colour
patches were not different between sexes. Achromatic
contrast in the white patch, red patch and dorsum
were significantly related to SVL, but this relationship
did not exist for any other colour patches (Supporting
Information, Table S2).
Sexual dimorphiSm oF the diSplay oF the
Frillneck lizard
We found a significant interaction effect between
SVL and sex for frill area (β = 18.566, SE = 4.994,
t = 3.717, P = 0.006). Male frill area was the largest
and increased more steeply with SVL (greater slope)
than both females (β = −26.508, SE = 13.061, d.f. = 42,
t = −2.029, P = 0.049) and juveniles (β = −26.060,
SE = 12.345, d.f. = 42, t = −2.111, P = 0.041; Fig. 5) .
Females had larger frills than juveniles; however,
the relationship between SVL and frill area was
not significantly different (β = −0.448, SE = 5.629,
t = −0.080, P = 0.937; Fig. 5).
Figure 1. The antipredator behavioural sequence of 52
frillneck lizards in response to the presentation of a model
raptor predator, showing the frequencies of transitions
after each behaviour. Behaviours are inside the boxes, the
proportions are adjacent to the arrows, and the dashed line
denotes a non-overlapping arrow. Flee (F) and stationary (S)
were considered endpoints of the sequence. Other associated
behaviours included tail lashes and hissing, although these
were not easily discernible in the behavioural assays.
Downloaded from https://academic.oup.com/biolinnean/advance-article-abstract/doi/10.1093/biolinnean/blz176/5679583 by University of Stellenbosch user on 19 December 2019
8 PEREZ-MARTINEZ ET AL.
© 2019 e Linnean Society of London, Biological Journal of the Linnean Society, 2019, XX, 1–14
We also found differences between sexes in display
behaviour. Males had a higher display score than
both females (β = −0.691, SE = 0.295, d.f. = 48,
t = −2.342, P = 0.023) and juveniles (β = −0.676,
SE = 0.321, d.f. = 48, t = −2.106, P = 0.04), but we
found no significant difference between females and
juveniles (β = −0.016, SE = 0.289, d.f. = 48, t = −0.054,
P = 0.958).
DISCUSSION
In accordance with our hypothesis, frillneck lizard
antipredator behaviour was consistent with deimatic
display theory. Frillneck lizards erected their frills
in response to the model avian predator significantly
more often than to the control pole. In response to
a predatory threat in controlled conditions, lizards
Figure 2. Mean spectra ± SE of four colour patches of 52 frillneck lizards and one of the bark background: dorsum (black
line and shading), represented by three regions on the lizard’s body and three regions on the edge of the frill; red–orange
patch (red line and shading); white patch (green line and shading); palate (purple line and shading); and eucalypt bark
(brown line and shading).
Downloaded from https://academic.oup.com/biolinnean/advance-article-abstract/doi/10.1093/biolinnean/blz176/5679583 by University of Stellenbosch user on 19 December 2019
FRILLNECK LIZARD ANTIPREDATOR BEHAVIOUR 9
© 2019 e Linnean Society of London, Biological Journal of the Linnean Society, 2019, XX, 1–14
Figure 3. Means ± 95% confidence intervals of chromatic (square) and achromatic (circle) contrasts of four colour patches
against the dorsum of the lizard in adults (purple) and juveniles (grey) in units of just-noticeable differences (JNDs) using a
raptor visual system. The vertical bold line demarcates the resting (left) and display (right) states, and the horizontal dotted
line indicates a discrimination threshold value of one for the JNDs (Siddiqi et al., 2004).
Figure 4. Means ± 95% confidence intervals of chromatic (square) and achromatic (circle) contrasts of the white patch
and palate against the dorsum of the lizard in males (blue), females (red) and juveniles (grey) in units of just-noticeable
differences (JNDs) using a raptor visual system. The horizontal dotted line indicates a discrimination threshold value of
one for the JNDs (Siddiqi et al., 2004).
Downloaded from https://academic.oup.com/biolinnean/advance-article-abstract/doi/10.1093/biolinnean/blz176/5679583 by University of Stellenbosch user on 19 December 2019
10 PEREZ-MARTINEZ ET AL.
© 2019 e Linnean Society of London, Biological Journal of the Linnean Society, 2019, XX, 1–14
exhibited a marked transition from a cryptic, resting
state to an active, display state. During rest, the lizard
remained stationary, with concealed colour patches.
In the display state, lizards erected the frill in concert
with colour patches on the frill and palate, which are
highly conspicuous to a raptor visual system. Frill
erection was accompanied by behaviours (lunging,
swaying, hissing and tail-whipping) oriented at the
predatory threat. Warning signals can often involve
multiple components and different sensory modalities
(Rowe & Guilford, 1999), thereby maximizing the
effect. Likewise, in a deimatic display, the abrupt
presentation of a visual stimulus may be concurrent
with movements, sounds or chemicals that amplify
its effect (Vallin et al., 2005; Umbers & Mappes, 2015;
Umbers et al., 2017).
Erecting the frill considerably increases the apparent
size of the lizard from the perspective of a predator; the
structure spans up to six times the width of the head in
adult males. Due to frill size alone, the misjudgement
of prey size may stall or halt the predatory sequence.
Moreover, the frill itself may serve as a diversion from
the body; it is fragile and susceptible to tears and
damage; therefore, attacks concentrated on the frill
would be less likely to incapacitate or gravely injure
the lizard.
Frill displays were triggered by the introduction
of the model raptor into the visual field of the lizard.
In all frillneck lizards that performed the display,
the behaviour was initiated simultaneously with the
appearance of the model avian predator over the arena
(within 0.20 s from the moment the model bird was
presented). Lizards are likely to deploy the frill during
the final stages of a predatory encounter. A late-stage
encounter is likely to have a greater impact on the
sensory system of a predator by virtue of proximity.
Behavioural observations by Shine (1990) and field
observations by C.A.P.-M. support the hypothesis
that lizards primarily initiate their displays at close
range. However, frillneck lizards may also display
from a distance of 10–50 m; these cases are uncommon
and have mainly been noted in response to a vehicle
(Shine, 1990). Perhaps, in these cases, the magnitude
of the threat triggers the display, because vehicles are
large, fast-moving and loud.
In addition to behaviour, colouration of the frillneck
lizard transitions from a cryptic to conspicuous state
during their display. The brown dorsum of the lizard
is distinguishable from the bark background; however,
given that the lizard remains motionless during the
resting state it is likely to be relatively camouflaged
to an avian predator. The erection of the frill exposes
patches on the frill and palate, all of which are highly
conspicuous to a raptor visual system. The spectral
reflectance of the palate revealed a strong ultraviolet
(UV) component. In northern blue-tongued skinks
(Tiliqua scinoides intermedia Mitchell, 1955), the
tongue is UV, conspicuous and thought to be a key
component of a deimatic display (Badiane et al., 2018).
In frillneck lizards, the palate is always exposed
Figure 5. Frill area in relationship to snout–vent length (SVL) across males (blue), females (red) and juveniles (grey), with
superimposed fitted lines over data points and shaded 95% confidence intervals.
Downloaded from https://academic.oup.com/biolinnean/advance-article-abstract/doi/10.1093/biolinnean/blz176/5679583 by University of Stellenbosch user on 19 December 2019
FRILLNECK LIZARD ANTIPREDATOR BEHAVIOUR 11
© 2019 e Linnean Society of London, Biological Journal of the Linnean Society, 2019, XX, 1–14
during displays, and we suggest that the UV-reflective
mouth acts as an additional component of the display,
amplifying the effect. Lastly, the two white patches
strongly contrast with the surrounding red, orange and
black colouration. Given the large frill as a backdrop,
we speculate that the white patches might convey
the appearance of eyespots to a predator. Overall, the
conspicuousness of the colouration of the frill adds
to the effect of the display by the frillneck lizard and
increases the potential of the display to elicit a startle
response in a predator.
Frillneck lizards are quick and agile. On the contrary,
most animals that perform deimatic displays are
relatively slow-moving, at least to the extent that they
would be unlikely to escape a predator once they are
confronted, such as the blue-tongued skink (Badiane
et al., 2018). When possible, we expect frillneck lizards
to flee and remain hidden from predators as their first
line of defense, only erecting the frill when cornered or
when escape seems improbable. Our enclosed arena did
not allow lizards this possibility, but we found that 17
of the 38 lizards that erected their frills subsequently
fled from the model until they made contact with
a side of the enclosure. It is possible that if assays
were performed in a larger enclosure, we might have
observed a higher proportion of lizards that fled after
displaying and, possibly, a higher proportion of lizards
that fled initially.
The toad-headed agama is another fast-moving
agamid whose antipredator behaviour conforms
to deimatic display theory (MJW, unpublished
observations). This species runs from an approaching
predator, and displays its head flaps only when it is
‘ambushed’ and restrained by a model predator (e.g.
the subjugation phase, which is when a predator bites
the animal; Vermeij, 1982). We speculate that early
in the evolution of the frill, frillneck lizards might
have had similar timing of the display, in which a
small proto-frill would be most effective as a predator
deterrent when a predator comes into contact with
the lizard. As larger frills were selected for, the timing
of their display behaviour might have shifted as the
structure became effective in predator deterrence from
a distance. This shift would allow the lizard to retain
the ability to startle predators while lowering the
probability of physical harm.
Juvenile frillneck lizards showed the greatest
conspicuousness of both chromatic and achromatic
contrast to both an avian and a lizard receiver. This
suggests that survival of smaller, younger and more
vulnerable dragons might rely on more conspicuous
display colouration. The predators differ between life
stages for many reptiles, because of a dramatic change
in body size during growth or a change in habitat
use (Irschick et al., 2000; Keren-Rotem et al., 2006;
Llewelyn et al., 2012; Purwandana et al., 2016). If frill
conspicuousness changes throughout the lifetime of an
individual, the question arises as to whether juveniles
face a different suite of predators, or predation pressure,
compared with adults. For example, because juveniles
have colour patches with greater conspicuousness to a
lizard receiver than adults, it is possible that juveniles
are at a higher risk from predation by lizard predators,
including conspecifics and goannas (e.g. Varanus
panoptes Storr, 1980). Juvenile frillneck lizards also
inhabit smaller trees and are more likely to perch in
dense understorey (C.A.P.-M., personal observation),
which might predispose them to different predation
pressure. This may imply that the targeted receiver
of the deimatic display by the frillneck lizard might
change as the animal matures. In the juvenile life-
stage, our findings suggest that the display might
target both avian and lizard predators. Future
research on deimatic displays should consider how it
might change throughout the lifetime of an individual.
Male competition for access to females and/or
resources can also drive the evolution of conspicuous
traits, resulting in sexual dimorphism (Andersson,
1982; Berglund et al., 1996; Macedonia et al., 2002).
Exaggeration of male traits, particularly sexual
dichromatism, is common in lizards (Stuart-Fox
& Ord, 2004) and can be used as an intraspecific
signal to assess fighting ability without escalation
into physical combat (e.g. Whiting et al., 2003, 2006,
2015). Frill colour has been proposed as a signal
used in territorial displays between males (Shine,
1990; Christian et al., 1995). In our study and others,
males have been documented to have larger frills
than females (Christian et al., 1995), although Shine
(1990) found no evidence of sexual dimorphism
in frill size. If the frill is a rare structure with a
dual role in deimatism and intrasexual selection,
we would expect the frill to have differences in
morphology and/or conspicuousness of colour
patches to a lizard visual system, with males having
larger and more conspicuous frills (e.g. Whiting
et al., 2003; Stuart-Fox & Ord, 2004). In line with
this hypothesis, our data showed that males had
larger frills than females and juveniles, along with
a disproportionate increase in frill area in relation
to SVL (Supporting Information, Table S2). Males
also had a significantly greater display score and
displayed their frills more vigorously than females.
However, male and female frills did not differ in
their colouration according to a lizard visual system,
with the exception of dorsum achromatic contrast
and palate chromatic contrast (Table 2). Therefore,
sexual selection might operate primarily on frill
size and display behaviour as opposed to frill colour.
Exactly how the frill is potentially used to signal
resource-holding potential, male quality or fighting
ability is a fruitful avenue for future research.
Downloaded from https://academic.oup.com/biolinnean/advance-article-abstract/doi/10.1093/biolinnean/blz176/5679583 by University of Stellenbosch user on 19 December 2019
12 PEREZ-MARTINEZ ET AL.
© 2019 e Linnean Society of London, Biological Journal of the Linnean Society, 2019, XX, 1–14
Deimatic displays are captivating examples of
antipredator behaviours, yet much empirical work
remains to be done to understand their role in the
survival of animals. Our results establish that the use
of the frill by Chlamydosaurus kingii in antipredator
behaviour conforms to the predictions of deimatic
display theory. The display is momentary, transient
and highly conspicuous, accompanied by behaviours
that amplify its effect. To our knowledge, there are
exceedingly few documented observations of frill
erection in response to a predator in situ (an exception
is the head flaps in the toad-headed agama). We
presume that the highly conspicuous frill elicits a
reflexive response in an avian predator (and perhaps
other types of predators), although this remains to
be tested. Frillneck lizard frill colouration was more
conspicuous in juveniles, potentially reflecting a
change in the nature of the deimatic display by this
species across ontogeny. Supporting this hypothesis,
the sexual size dimorphism in the frill and behavioural
difference in the display (males having larger frills and
more vigorous displays) hints at a possible dual role in
antipredator behaviour and sexual selection. If this is
true, it will be a rare instance of an antipredator display
being co-opted for another function (social signalling).
We recommend that future research should examine
the behavioural responses of predators to frillneck
lizard displays, in addition to staged dyadic contests to
test the role of the frill in contest competition.
ACKNOWLEDGEMENTS
We would like to thank the Tropical Ecology Research
Facility of the University of Sydney, Rick Shine, Greg
Brown and Thomas Madsen for general support and
advice in conducting fieldwork and running behavioural
assays, and the Fogg Dam Rangers for offering their
support during the project. Thanks to Almut Kelber
for her correspondence and advice regarding the visual
modelling component of our analyses, and thanks to
James Baxter-Gilbert, Greg Clarke, Maddie Sanders,
Roshmi Sarma and Cathy Shilton for logistical support
in the field. We are also grateful to three anonymous
reviewers for improving this manuscript. This work
was funded by the Alex G. Booth Fellowship of Harvard
College awarded to C.A.P.-M., and an Endeavour
Postdoctoral Fellowship and Claude Leon Foundation
Postdoctoral Fellowship awarded to J.L.R. The authors
declare no conflicts of interest.
REFERENCES
Andersson M. 1982. Sexual selection, natural selection and
quality advertisement. Biological Journal of the Linnean
Society 17: 375–393.
Bacchus J. 1939. Notable Australian lizards. Walkabout 5:
53–60.
Badiane A, Carazo P, Price-Rees SJ, Ferrando-Bernal M,
Whiting MJ. 2018. Why blue tongue? a potential UV-based
deimatic display in a lizard. Behavioral Ecology and
Sociobiology 72: 104.
Berglund A, Bisazza A, Pilastro A. 1996. Armaments and
ornaments: an evolutionary explanation of traits of dual
utility. Biological Journal of the Linnean Society 58: 385–399.
Bureau of Meteorology. 2018. Average annual, seasonal
and monthly rainfall. Available at: http://www.bom.gov.au/
climate/data/index.shtml
Christian KA, Bedford GS. 1995. Seasonal changes in
thermoregulation by the frillneck lizard, Chlamydosaurus
kingii, in tropical Australia. Ecology 76: 124–132.
Christian KA, Griffiths A, Bedford G, Jenkin G. 1999.
Androgen concentrations and behavior of frillneck lizards
(Chlamydosaurus kingii). Journal of Herpetology 33: 12–17.
Cogger H. 2002. Reptiles of Australia. Sydney: Reed New
Holland.
De Vis CW. 1883. Myology of Chlamydosaurus kingii.
Proceedings of the Linnean Society of New South Wales 1:
300–320.
Endler J. 1978. A predator’s view of animal color patterns.
Evolutionary Biology 11: 319–364.
Endler JA, Mappes J. 2017. The current and future state of
animal coloration research. Philosophical Transactions of the
Royal Society B: Biological Sciences 372: 20160352.
Fenner C. 1933. Bunyips and billabongs. Sydney: Angus &
Robertson.
Fleishman LJ, Loew ER, Whiting MJ. 2011. High
sensitivity to short wavelengths in a lizard and implications
for understanding the evolution of visual systems in lizards.
Proceedings of the Royal Society B: Biological Sciences 278:
2891–2899.
Friard O, Gamba M. 2016. BORIS: a free, versatile open-
source event-logging software for video/audio coding and
live observations. Methods in Ecology and Evolution 7:
1325–1330.
Frith C, Frith D. 1987. Australian tropical reptiles and frogs.
Townsville: Tropical Australia Graphics.
Griffiths AD, Christian KA. 1996. Diet and habitat use
of frillneck lizards in a seasonal tropical environment.
Oecologia 106: 39–48.
Holmes GG, Delferrière E, Rowe C, Troscianko J,
Skelhorn J. 2018. Testing the feasibility of the startle-first
route to deimatism. Scientific Reports 8: 10737.
Irschick DJ, Macrini TE, Koruba S, Forman J. 2000.
Ontogenetic differences in morphology, habitat use, behavior,
and sprinting capacity in two West Indian Anolis lizards.
Journal of Herpetology 34: 444–451.
Kemp DJ, Herberstein ME, Fleishman LJ, Endler JA,
Bennett ATD, Dyer AG, Hart NS, Marshall J,
Downloaded from https://academic.oup.com/biolinnean/advance-article-abstract/doi/10.1093/biolinnean/blz176/5679583 by University of Stellenbosch user on 19 December 2019
FRILLNECK LIZARD ANTIPREDATOR BEHAVIOUR 13
© 2019 e Linnean Society of London, Biological Journal of the Linnean Society, 2019, XX, 1–14
Whiting MJ. 2015. An integrative framework for the
appraisal of coloration in nature. The American Naturalist
185: 705–724.
Keren-Rotem T, Bouskila A, Geffen E. 2006. Ontogenetic
habitat shift and risk of cannibalism in the common
chameleon (Chamaeleo chamaeleon). Behavioral Ecology
and Sociobiology 59: 723–731.
Lenth RV. 2016. Least-squares means: the R package lsmeans.
Journal of Statistical Software 69: 1–33.
Lind O, Mitkus M, Olsson P, Kelber A. 2013. Ultraviolet
sensitivity and colour vision in raptor foraging. Journal of
Experimental Biology 216: 1819–1826.
Llewelyn J, Bull K, Schwarzkopf L, Alford RA, Shine R.
2012. Ontogenetic shifts in a prey’s chemical defences
influence feeding response of a snake predator. Oecologia
169: 965–973.
Loeffler-Henry K, Kang C, Yip Y, Caro T, Sherratt TN.
2018. Flash behavior increases prey survival. Behavioral
Ecology 29: 528–533.
MacDougall A, Dawkins MS. 1998. Predator discrimination
error and the benefits of Müllerian mimicry. Animal
Behaviour 55: 1281–1288.
Macedonia JM, Brandt Y, Clark DL. 2002. Sexual
dichromatism and differential conspicuousness in two
populations of the common collared lizard (Crotaphytus
collaris) from Utah and New Mexico, USA. Biological
Journal of the Linnean Society 77: 67–85.
Maia R, Eliason CM, Bitton P-P, Doucet SM, Shawkey MD.
2013. pavo: an R package for the analysis, visualization
and organization of spectral data. Methods in Ecology and
Evolution 4: 906–913.
Maldonado H. 1970. The deimatic reaction in the
praying mantis Stagmatoptera biocellata. Zeitschrift für
Vergleichende Physiologie 68: 60–71.
Mappes J, Marples N, Endler JA. 2005. The complex
business of survival by aposematism. Trends in Ecology &
Evolution 20: 598–603.
McLean CA, Lutz A, Rankin KJ, Elliott A, Moussalli A,
Stuart-Fox D. 2019. Red carotenoids and associated gene
expression explain colour variation in frillneck lizards.
Proceedings of the Royal Society B: Biological Sciences 286:
20191172.
Nelson XJ. 2012. A predator’s perspective of the accuracy of
ant mimicry in spiders. Psyche 2012: 168549.
Purwandana, D, Ariefiandy A, Imansyah MJ, Seno A,
Ciofi C, Letnic M, Jessop TS. 2016. Ecological allometries
and niche use dynamics across Komodo dragon ontogeny.
The Science of Nature 103: 27.
R Core Team. 2018. R: a language and environment for
statistical computing. Vienna: R Foundation for Statistical
Computing.
Riley JL, Noble DWA, Bryne RW, Whiting MJ. 2017. Early
social environment influences the behaviour of a family-
living lizard. Royal Society Open Science 4: 161082.
Rowe C, Guilford T. 1999. The evolution of multimodal
warning displays. Evolutionary Ecology 13: 655–671.
Rowe C, Haplin C. 2013. Why are warning displays
multimodal? Behavioral Ecology and Sociobiology 67:
1425–1439.
Saville-Kent W. 1896. The frilled lizard: “Chlamydosaurus
kingi”. Nature 53: 395–398.
Sergo D, Shine R. 2015. Snakes for lunch: bird predation
on reptiles in a tropical floodplain. Australian Zoologist 37:
311–320.
Shine, R. 1990. Function and evolution of the frill of the
frillneck lizard, Chlamydosaurus kingii (Sauria: Agamidae).
Biological Journal of the Linnean Society 40: 11–20.
Siddiqi A, Cronon TW, Loew ER, Vorobyev M, Summers K.
2004. Interspecific and intraspecific views of color signals in
the strawberry poison frog Dendrobates pumilio. The Journal
of Experimental Biology 207: 2471–2485.
Stefanescu C, Berger VW, Hershberger SL. 2005. Yates’
correction. In Encyclopedia of statistics in behavioral science.
American Cancer Society.
Stevens M. 2005. The role of eyespots as anti-predator
mechanisms, principally demonstrated in the lepidoptera.
Biological Reviews 80: 573–588.
Stevens M, Merilaita S. 2009. Animal camouflage: current
issues and new perspectives. Philosophical Transactions of
the Royal Society B: Biological Sciences 364: 423–427.
Stevens M, Ruxton GD. 2012. Linking the evolution and
form of warning coloration in nature. Proceedings of the
Royal Society B: Biological Sciences 279: 417–426.
Stuart-Fox D, Adnan M, Whiting MJ. 2008. Predator-specific
camouflage in chameleons. Biology Letters 4: 326–329.
Stuart-Fox DM, Ord TJ. 2004. Sexual selection, natural
selection and the evolution of dimorphic coloration and
ornamentation in agamid lizards. Proceedings of the Royal
Society B: Biological Sciences 271: 2249–2255.
Stuart-Fox DM, Whiting MJ, Moussalli A. 2006. Camouflage
and colour change: antipredator responses to bird and snake
predators across multiple populations in a dwarf chameleon.
Biological Journal of the Linnean Society 88: 437–446.
Ujvari B, Fisher P, Rydell J, Wahlgren R, Wright B,
Madsen T. 2015. Population demography of frillneck lizards
(Chlamydosaurus kingii, Gray 1825) in the wet-dry tropics of
Australia. Australian Ecology 40: 60–66.
Umbers KDL, De Bona S, White TE, Lehtonen J, Mappes J,
Endler JA. 2017. Deimatism: a neglected component of
antipredator defence. Biology Letters 13: 20160936.
Umbers KDL, Lehtonen J, Mappes J. 2015. Deimatic
displays. Current Biology: CB 25: R58–R59.
Umbers KDL, Mappes J. 2015. Postattack deimatic display
in the mountain katydid, Acripeza reticulata. Animal
Behaviour 100: 68–73.
Umbers KDL, Mappes J. 2016. Towards a tractable working
hypothesis for deimatic displays. Animal Behaviour 113: e5–e7.
Umeton D, Read JCA, Rowe C. 2017. Unravelling the
illusion of flicker fusion. Biology Letters 13: 20160831.
Vallin A, Jakobsson S, Lind J, Wiklund C. 2005. Prey
survival by predator intimidation: an experimental study of
peacock butterfly defence against blue tits. Proceedings of
the Royal Society B: Biological Sciences 272: 1203–1207.
Vermeij GJ. 1982. Unsuccessful predation and evolution. The
American Naturalist 120: 701–720.
Vorobyev M, Osorio D. 1998. Receptor noise as a determinant
of colour thresholds. Proceedings of the Royal Society of
London B: Biological Sciences 265: 351–358.
Downloaded from https://academic.oup.com/biolinnean/advance-article-abstract/doi/10.1093/biolinnean/blz176/5679583 by University of Stellenbosch user on 19 December 2019
14 PEREZ-MARTINEZ ET AL.
© 2019 e Linnean Society of London, Biological Journal of the Linnean Society, 2019, XX, 1–14
Wedmann S. 2010. A brief review of the fossil history of plant
masquerade by insects. Palaeontographica Abteilung B 283:
175–182.
Whiting MJ, Nagy KA, Bateman PW. 2003. Evolution and
maintenance of social status signalling badges: experimental
manipulations in lizards. In Fox SF, McCoy JK, Baird TA,
eds. Lizard social behavior. Baltimore: Johns Hopkins
University Press, 47–82.
Whiting MJ, Noble DWA, Somaweera R. 2015. Sexual
dimorphism in conspicuousness and ornamentation in the
enigmatic leaf-nosed lizard Ceratophora tennentii from
Sri Lanka. Biological Journal of the Linnean Society 116:
614–625.
Whiting MJ, Stuart-Fox DM, O’Connor D, Firth D,
Bennett NC, Blomberg SP. 2006. Ultraviolet signals
ultra-aggression in a lizard. Animal Behaviour 72:
353–363.
Worrell E. 1963. Reptiles of Australia. Sydney: Angus &
Robertson.
Zuur AF, Ieno EN, Elphick CS. 2010. A protocol for data
exploration to avoid common statistical problems. Methods
in Ecology and Evolution 1: 3–14.
SUPPORTING INFORMATION
Additional supporting information may be found in the online version of this article at the publisher’s website:
Supplementary methods. Sexing wild-caught lizards, display score summary and visual modelling.
Figure S1. Diagram of measurements taken to calculate frill area. Yellow lines (2–4) correspond to hyoid
cartilaginous segments. White lines (5–25) are used to represent the remainder of half of the frill.
Figure S2. Photographic representation of the resting and display states of the frillneck lizard, with corresponding
regions where we took spectrophotometric measurements.
Figure S3. Graphical representation of principal components analysis (PCA) for head morphology of male (blue),
female (red) and juvenile (grey) frillneck lizards. Circles represent 95% confidence intervals. The PCA included
head height (in millimetres), head width (in millimetres) and head length (in millimetres) measurements for 52
lizards. All morphological measurements were log10-transformed.
Table S1. Output of linear models exploring whether sex and snout–vent length are predictors of chromatic and
achromatic contrasts of colour patches according to the visual system of: (1) the common buzzard (Buteo buteo),
representing an avian predator; and (2) the ornate-crevice dragon (Ctenophorus ornatus), representing a frillneck
lizard conspecific. Bold P-values indicate significance at a level of α = 0.05, and reference levels of the categorical
variable sex are presented in parentheses after its name.
Table S2. Output of a linear model exploring whether sex or snout–vent length is a predictor of frill area. Bold
P-values indicate significance at a level of α = 0.05, and reference levels of the categorical variable sex are presented
in parentheses after its name. Post hoc multiple comparisons between effects can be found in the main text.
Table S3. Component loadings for first (PC1) and second (PC2) principal components from our principal
components analysis (PCA) for head morphology of frillneck lizards. The PCA included log10-transformed head
height (in millimetres), head width (in millimetres) and head length (in millimetres) measurements for 52 lizards.
We also present the standard deviation of PC1 and PC2, in addition to the proportion of variance they explain in
our data. See the Supporting Information (Fig. S3) for a graphical representation of these data.
Table S4. Component loadings for the first principal component (PC1) from our principal components analysis
(PCA) for display behavioural scores. The PCA included the behavioural categories used in the formula for
behavioural score in the main text in the main text, with the exception of flee (R) for 52 lizards. We also present
the standard deviation of PC1 and the proportion of variance it explains in our data.
SHARED DATA
All data and R code from this study can be accessed at doi:10.17605/OSF.IO/98F23.
Downloaded from https://academic.oup.com/biolinnean/advance-article-abstract/doi/10.1093/biolinnean/blz176/5679583 by University of Stellenbosch user on 19 December 2019
