The Science of Nature (2023) 110:23

https://doi.org/10.1007/s00114-023-01849-6

ORIGINAL ARTICLE

Visual antipredator effects of web flexing in an orb web spider,

with special reference to web decorations
Luis E. Robledo‑Ospina1,2 · Nathan Morehouse3 · Federico Escobar1 · Horacio Tapia‑McClung4 ·
Ajay Narendra5 · Dinesh Rao2

Received: 27 January 2023 / Revised: 5 May 2023 / Accepted: 9 May 2023 / Published online: 23 May 2023
© The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2023

Abstract
Some visual antipredator strategies involve the rapid movement of highly contrasting body patterns to frighten or confuse
the predator. Bright body colouration, however, can also be detected by potential predators and used as a cue. Among spi-
ders, Argiope spp. are usually brightly coloured but they are not a common item in the diet of araneophagic wasps. When
disturbed, Argiope executes a web-flexing behaviour in which they move rapidly and may be perceived as if they move
backwards and towards an observer in front of the web. We studied the mechanisms underlying web-flexing behaviour as
a defensive strategy. Using multispectral images and high-speed videos with deep-learning-based tracking techniques, we
evaluated body colouration, body pattern, and spider kinematics from the perspective of a potential wasp predator. We show
that the spider’s abdomen is conspicuous, with a disruptive colouration pattern. We found that the body outline of spiders
with web decorations was harder to detect when compared to spiders without decorations. The abdomen was also the body
part that moved fastest, and its motion was composed mainly of translational (vertical) vectors in the potential predator’s
optical flow. In addition, with high contrast colouration, the spider’s movement might be perceived as a sudden change in
body size (looming effect) as perceived by the predator. These effects alongside the other visual cues may confuse potential
wasp predators by breaking the spider body outline and affecting the wasp’s flight manoeuvre, thereby deterring the wasp
from executing the final attack.

Keywords Deimatic displays · High-contrast visual cues · Secondary defensive strategies · Visual ecology

Introduction

Animals use diverse antipredator sensory defences to pre-

vent, interrupt, and/or stop predator attack (Ruxton et al.
2018) depending on the stage of the predator-prey attack
Communicated by: Matjaž Gregorič sequence when the interaction occurs (Endler 1991). Prey
often use crypsis and immobility to avoid detection, but if
* Dinesh Rao those defences fail, other strategies are needed to avoid being
[email protected] captured and killed (Heathcote et al. 2020). Most animals can-
1
Red de Ecoetología, Instituto de Ecología, A.C., Xalapa,
not remain still indefinitely, and any subsequent motion ren-
Veracruz, México ders them more vulnerable to detection (Hughes et al. 2014,
2
Instituto de Biotecnología y Ecología Aplicada, Universidad
Umeton et al. 2017, Umeton et al. 2019, Tan and Elgar 2021).
Veracruzana, Xalapa, Veracruz, México When a prey detects a predator, secondary defence strategies
3
Department of Biological Sciences, University of Cincinnati,
are triggered to prevent capture, often depriving predators of
Cincinnati, OH, USA information or providing confusing cues (Caro 2014).
4
Instituto de Investigaciones en Inteligencia Artificial,
Secondary strategies such as motion dazzle and dei-
Universidad Veracruzana, Xalapa, Veracruz, México matic (startle) displays are part of the diverse repertoire of
5
School of Natural Sciences, Macquarie University, Sydney,
antipredator strategies in animals. Motion dazzle involves
NSW 2109, Australia the use of high contrast patterning that hinders estimation

13
Vol.:(0123456789)
23 Page 2 of 13 The Science of Nature (2023) 110:23

of prey trajectory and speed when the target is in motion Orb-web spiders use secondary defence strategies such as
(Scott-Samuel et al. 2011, Stevens and Merilaita 2011). In dropping to the ground, changing sides on the web, and/or
contrast, deimatic displays involves the revealing of previ- web vibration/flexing to avoid capture (Cloudsley-Thompson
ously concealed conspicuous colouration patterns along with 1995, Pekár 2014, Gawryszewski 2017). Web-flexing occurs
sudden movements directed towards the attacker (Umbers when the rapid extension-retraction of the spider legs cause
et al. 2015, Drinkwater et al. 2022). Through the use of con- motion along the web’s short axis, also perceived as mov-
spicuous colouration and prey motion, both strategies are ing backwards and forward (Tolbert 1975). This behaviour
exploiting the three primary parameters of vision in ani- has been recorded in numerous orb-web spiders and may
mals: spectral sensitivity (colour perception), spatial resolu- be more effective on webs with decorations (stabilimenta
tion (visual acuity), and temporal resolution (Cronin et al. or conspicuous silk structures attached to the web) because
2014). Hence, while dazzling signals deceive the predator’s such decorations might help to blur the spider’s outline
estimation of the prey’s direction and speed (Hughes et al. and conceal its form (Cloudsley-Thompson 1995). How-
2017), deimatism involves behaviours that give rise to a ever, there is no conclusive evidence of it, and spiders with
sudden transition in sensory input, causing the predator to decorations may also use other antipredator tactics such as
respond reflexively (Umbers et al. 2017). Both cases might crypsis and web shaking (Bateman and Fleming 2013).
produce optical illusions (i.e., errors of perception) through Spiders are thought to use web decorations (see also
body colouration, movement, or manipulation of the envi- Eberhard, 2020) to either attract prey (Li et al. 2004, Cheng
ronment (Kelley and Kelley 2014). These optical illusions and Tso 2007) or hide from predators (Gonzaga and Vas-
may thus have evolved to hamper predation in different prey concellos-Neto 2005, Nakata 2009, Soley 2019, Wang
species (Kodandaramaiah et al. 2020, Valkonen et al. 2020). et al. 2021a), and even to protect their web from mechani-
Additionally, looming visual stimuli provide a visual cue cal damage (Blackledge and Wenzel 1999). In spiders from
of an approaching object (Spano et al. 2012), and observer the Argiope genus, decorations are detectable to both prey
animals innately respond to these cues to avoid a collision and predators (Robinson and Robinson 1970, Eisner and
(Peron and Gabbiani 2009, Spano et al. 2012, Tyll et al. Nowicki 1983, Bruce et al. 2005, Rao et al. 2008). However,
2013, Yilmaz and Meister 2013). This behaviour has been despite their conspicuousness and diurnal habitus, Argiope
addressed mostly from the perspective of individuals, often spp. spiders also have secondary defences (Tolbert 1975,
prey species, that respond by attempting to avoid the threat Jackson 1992, Blackledge and Pickett 2000).
of rapidly approaching objects (Temizer et al. 2015, Vag- In Argiope spiders, web-flexing may deter predators by
noni et al. 2015, Shragai et al. 2017, Donohue et al. 2022). blurring the spider’s outline (Robinson and Robinson 1970)
However, such looming stimuli may also work as a defen- and temporarily concealing its exact location (Tolbert 1975).
sive strategy against predators. A looming cue may be more However, the visually mediated functions of web decorations
effective when coupled with high contrast patterns, like the have not been explored from the perspective of predator vis-
stripes in moving zebras that may hinder the approach and ual systems (Théry and Casas 2009). Consequently, one of
landing stages of blood-sucking flies (Caro et al. 2019). In the outstanding questions regarding orb-weaver antipreda-
flying insects, the processing of looming signals is crucial tor responses is what mechanisms deter predators based
for navigation and object avoidance during the flight (Mui- on insight from predator perception (Umbers et al. 2017).
jres et al. 2014, Ache et al. 2019). Furthermore, moving high Accordingly, we studied the complex antipredator displays
contrast patterns may produce errors in motion detection of Argiope spp. First, we hypothesised that different body
mechanisms by mismatching local image contours (How and parts may have different visual contrasts (chromaticity and
Zanker 2014). luminance), making them useful in antipredator strategies
A good model to evaluate visual signalling in a movement that require high visual contrast. Second, because of the
triggered antipredator behaviour is the interaction between moving scene dynamism and the subsequent variation of
web-building spiders and araneophagic wasps. These wasps the different scene elements presented by web-flexing behav-
use both chemical and optical information when hunting for iours under natural circumstances (i.e., the combined scene
prey (Uma and Weiss 2010). These wasps lay significant involving movements of the spider, web decorations, and
emphasis in learning and using visual information in their background), predator perception and associated responses
natural habitat (e.g. Eberhard 1970; Zeil 1993). In addition, will depend on contrasts between these elements as viewed
there is evidence that spider body colouration may play a from different distances over the course of an attack. We
critical role at the final stages of the detection and attack hypothesised that, when such a spatio-temporal analysis
sequence, making spiders with conspicuous colouration was performed, we would find that web-flexing presents a
more likely to be detected by predatory wasps (Robledo- looming stimulus to prospective predators, augmented by
Ospina et al. 2021). web decorations (when present). This would in turn elicit

13
The Science of Nature (2023) 110:23 Page 3 of 13 23

avoidance responses rather than attraction, despite the high wavelengths with high efficiency between 300 and 400 nm.
saliency of Argiope spiders when evaluated from a more All the images included a scale bar and two Zenith sintered
static perceptual paradigm. PTFE 70% and 10% grey standards. We used natural sun-
To investigate the mechanisms underlying the web-flex- light as a light source (between 1000 and 1500 h) and local
ing defensive behaviour in orb-weaver spiders, we measured natural vegetation as background at constant distance from
the body colouration, shape, and movement of an Argiope the spider (1 m). Using full-spectrum digital photography
aurantia spider as perceived by a potential wasp predator (Troscianko and Stevens 2015), we created multispectral
during web-flexing behaviour. Using full-spectrum digital images of the spider in its web, choosing as regions of inter-
photography, high-speed videography, and psychophysiolog- est (ROIs) the spider body (cephalothorax and abdomen),
ical models, we aimed to describe a behaviour that includes the web decoration, and the natural background.
not only chromatic, but also motion perception since this All the resulting multispectral images were analysed
type of sensory processing often underpins prey detection, using Quantitative Pattern Colour Analysis (QCPA), follow-
avoidance of predators and communication with conspecifics ing the methods of Van den Berg et al. (2019). This meth-
(Peters et al. 2002). odological framework uses the receptor noise-limited (RNL)
model proposed by Vorobyev and Osorio (1998) for colour
vision and a visual acuity correction approach that, in our
Materials and methods case, was based on Gaussian filters to simulate the receiver
spatial acuity and viewing distance.
Adult females of Argiope aurantia (Araneae: Araneidae) Regarding the visual system used for modelling, there
(Lucas 1833) often perform a web-flexing behaviour as is no information published about vision parameters like
an antipredator response when disturbed (Tolbert 1975; spectral sensitivity and spatial acuity for the araneopha-
Robledo, pers. obs.). We collected 14 adult female indi- gic mud-dauber wasp, Trypoxylon Richards (Crabronidae),
viduals from different locations in the city of Xalapa, Ver- which attack and prey on Argiope spiders (Robledo-Ospina
acruz, Mexico. We housed the spiders at the Universidad et al., 2021). We therefore used parameters from a related
Veracruzana in wooden frames (67*67*19 cm) with trans- species Philanthus triangulum Fabricius (Sphecidae), to cre-
parent Plexiglas covers. Spiders were fed every other day ate a putative model of an araneophagic wasp visual system.
with small-to-medium (~35 mm body length) grasshoppers This wasp is related to Trypoxylon wasps and both species
(Acrididae: Orthoptera) collected in the laboratory’s sur- are spider-hunting specialists (Robledo-Ospina et al. 2022).
roundings, and water was sprayed on the web during feeding. Philanthus triangulum has trichromatic vision which is
We established two main frameworks to evaluate the vis- typical of Hymenopterans, with sensitivity peaks at 344 nm
ual signals involved in web-flexing behaviour in A. aurantia (UV), 444 nm (SW), and 524 nm (MW, Peitsch et al. 1992).
as perceived by a hypothetical wasp predator, based on the For the RNL model, we used the Weber fraction (ω = 0.13)
measurement of image parameters like brightness, contrast, and the relative density for each receptor class (ni) used by
and spatial frequency composition, which may influence Defrize et al. (2010) based on Apis mellifera (1:0.471:4.412
motion perception (Egelhaaf et al. 2012). Based on these ratios for the UV:SW:MW receptors, respectively). Moreo-
measurements, (1) we estimated the detectability of the ver, we used the minimal resolvable angle for the visual acu-
immobile spider against their web decoration and the back- ity value reported for Bembix palmata Smith (Crabronidae)
ground at different distances using a simulated wasp visual of 1.22 cycles per degree (Feller et al. 2020).
system with both chromatic/achromatic contrast perception, In the QCPA plugin, we simulated the wasp visual acu-
combining these elements with visual acuity and edge sali- ity to eliminate details that the receiver visual system cannot
ency. (2) We then described and measured the kinematics of resolve (Caves and Johnsen 2017). Then, we used the RNL
the web-flexing behaviour, which allowed us to estimate the ranked filter to model the perception of spatial information in
perceptual changes of body shape and size during the spider our wasp visual system (Van der Berg et al. 2019) at different
web-flexing as perceived by a wasp. distances of observation, simulating an approaching flying
wasp (10, 5, and 2 cm). These distances were chosen based
Detectability of the spider in the web on a field study on a wasp’s behavioural response to a flower
dwelling predator (Rodríguez-Morales et al. 2021). We also
We photographed spiders using a modified Canon 7D cam- used a colour adjacency analysis (CAA) to measure the colour
era without the internal ultraviolet (UV) filter (LifeP​ixel.​ pattern properties in the scene (Endler 2012), and estimated
com), with a UV transmitting Nikkor EL 80 mm lens. We the colour diversity and pattern complexity with indices based
took two types of photos per spider: a visible-light photo on the Shannon diversity index. In these indices, colour pat-
using a Baader UV/IR Cut filter (400–700 nm), and a terns with uniform spacing and coverage would have a maxi-
UV photo taken with a Baader U filter that transmits UV mum value of 1 (Van der Berg et al. 2019).

13
23 Page 4 of 13 The Science of Nature (2023) 110:23

Because motion-detection models based on spatio-tempo- the body: one in the cephalothorax anterior end (Head), and
ral correlation depend on contrast (How and Zanker 2014), two in the terminal points of the abdomen’s longest axis (Abd
we evaluated the mean luminance (M∆SL) and chromatic P, Abd A for the posterior and anterior ends, respectively).
contrast (M∆S) of the spider (cephalothorax and abdomen Six points were selected in the femur’s mid-portion of the
apart) against its web decoration and background, weighted legs II to IV in both spider sides (left and right) because they
by the mean relative abundance of each colour pattern ele- are points with high contrast colouration, which favour the
ment combination in the scene (Van der Berg et al. 2019). tracking process and may be more salient from a wasp’s per-
We also estimated the standard deviation of the RNL lumi- spective. We used DeepLabCut (DLC, ver 1.9) (Mathis et al.
nance and chromaticity contrasts following Endler et al. 2018) to track the points on the spider. Using DLC, we trained
(2018). We used the MW photoreceptor values for the lumi- a Resnet-50 network. We used 20 randomly selected frames
nance channel because hymenopterans use this receptor for each from 14 videos to train the network (Supplementary
luminance perception (Théry et al. 2005). Material). We trimmed the videos to exclude frames where the
Finally, after considering the wasp’s visual acuity and the tuning fork was visible. The network was trained for 100,000
RNL ranked filter to the images, we used the “Local Edge iterations. After training, all videos were analysed for location
Intensity Analysis” (LEIA) to quantify edge properties in the and movement of the landmark points XY (converted to mm)
scene in terms of colour and luminance contrast in log-linear positions of the points. These location sequences were then
RNL colour space (Van der Berg et al. 2019). We used the extracted and analysed using Mathematica Ver 12 (Wolfram
measure of kurtosis for both chromatic and luminance con- Research Inc). During analysis, we measured and compared
trast as descriptors of edge intensities of the scene from this the point trajectories to assess each body part’s contribution
analysis. Thus, higher kurtosis values indicate more salient to the movement as perceived by an observer during the initial
patterns in the scene. flexing cycle. The cycle in the flexing behaviour was meas-
ured from the spider resting position (Rest), going backwards
Visual signalling in motion: kinematics and shape in a first push (Back), going forward (Front), and then recover-
changes ing the initial position.
To estimate perceived changes in body size during a cycle,
The web-flexing antipredator behaviours of Argiope spp. we used the statistical package concaveman (Gombin et al.
spiders have been studied from a behavioural ecology per- 2017) in R statistical software (R Development Core Team
spective (e.g. Cedhagen and Björklund 2007; Jackson et al. 2016) to calculate the area of the concave hull polygon that
1993; Li and Lee 2004; Robinson and Robinson 1970; contains the tracked point abovementioned and fitted to the
Walter 2019). Hence, based on these findings, we induced focal spider body part for every frame. Then, we estimated,
these behaviours in live spiders by stimulating the abdomen and compared the area for the three cycle reference positions:
(Hoffmaster 1982) through both high-frequency vibration rest, back, and front.
stimuli (440hz) using a tuning fork (Nakata 2009) close to
the spider’s body (< 2 cm), or by gently touching the abdo- Optical flow
men (Jackson 1992). Web-flexing behaviour was recorded
perpendicular to the web plane using a Chronos 1.4 high- In flight, wasps, like other flying insects, perceive spatial
speed camera (Kron Technologies Inc., Burnaby, Canada) information as an optic flow of moving traits (Egelhaaf
with a 12mm f/1.4 prime lens (Computar, NC, USA) film- 2023). Such real-time analysis of motion, both self-induced
ing at 500fps with a resolution of 1280 × 1024 pixels and and from other (moving) objects, is critical for flight control
saved in MP4 format. The first 5 s of the web-flexing (~2500 (Srinivasan 1992, Zeil 1993, Zeil 1997, Baird et al. 2011,
frames) were used as event sampling for subsequent analy- Cronin et al. 2014, Serres and Ruffier 2017, Chakravarthi
ses. We recorded web-flexing behaviour four times for each et al. 2018, Horridge 2019). Hence, we aimed to objectively
spider but tracked only the first cycle. visualise the magnitude and direction of pixel displacement
In combination with high contrast patterns, movement (Raudies 2013) involved in the optical flow vector field pro-
is a remarkable feature of this behaviour and measuring it duced by the spider when performing web-flexing. We used
objectively is essential to estimating how an observer may an optical flow algorithm through the ImageDisplacements
perceive it. According to most theoretical and empirical analy- function in Mathematica (ver 11), which generates a dense
ses of whole-scene motion, the input parameters critical to motion field by comparing the displacement in the horizon-
viewer perception are stimulus direction and speed (Peters tal and vertical axis of individual pixels across the frames of
et al. 2002). Consequently, we measured the kinematics of the the 5-s sampling video. This method allows us to have a rep-
behaviour by tracking the changes in the space of nine points resentation of the potential perception of web-flexing motion
in the spider’s body. These focal points were selected because components from the wasp’s perspective, which can inform
they represent the distal and mid portion of the medial axis of the possibility that such displays generate optical illusions

13
The Science of Nature (2023) 110:23 Page 5 of 13 23

Fig. 1  Adult female of Argiope aurantia with a typical linear decora- umn shows the results of the receptor noise limited method, perform-
tion in its web. A Note the white web decoration which is particularly ing pixel noise reduction after the acuity control based on Gaussian
large in this individual’s web. B False colour image simulating the filters while preserving chromatic and luminance edges, simulating
wasp Philanthus triangulum colour vision and created for visualisa- wasp spectral sensitivity (sensitivity peaks are UV = 344, SW = 444,
tion purposes by assigning blue, green, and red for the UV, SW, and and MW = 524; Weber fraction = 0.13) and visual acuity (1.22 cpd)
MW wasp photoreceptors, respectively. C Results of the quantita- of the wasp visual system at different distances; right column presents
tive colour pattern analysis (QCPA) of the spider as perceived by the the results of the achromatic edge intensity of the LEIA at different
wasp at different distances (10, 5, and 2 cm, respectively): left col- distances where different colours indicate different angle of contrast

for receivers (How and Zanker 2014). We therefore analysed the two cameras and reconstructed the 3D position using the
different moving elements in the scene and their contribu- pose3d application (Sheshadri et al. 2020) in Matlab 2020b
tion to the rotational (horizontal) and translational (vertical) (MATLAB, 2020). The X, Y, and Z coordinates were then used
components of optic flow fields. for further analysis. To estimate the extent of displacement of
each of the 9 points tracked, we used a convex hull method
(Mathematica, ver 11). We computed the volume of the convex
3D tracking
hull and used it as a proxy to determine differences in move-
ment of the points regarding changes in depth.
We reconstructed the flexing behaviour of a single spider in
3D. Web-flexing behaviour was triggered as detailed above
(see kinematics section in methods). We filmed the movement Statistical analysis
of the spider at 500 fps with two synchronised and calibrated
high-speed Chronos 1.4 cameras with 12mm f/1.4 prime lens. From the photographic and videographic data described above,
The cameras were oriented towards the spider at 45° and 60°, we measured the body area (as an estimation of spider size)
respectively. Only one cycle of the web-flexing behaviour (rest and the web decoration area using the ImageJ software (Sch-
position, backward; front, rest position) was used for this analy- neider et al. 2012). To evaluate if the web decoration area has
sis. We tracked 9 points selected according to the areas of maxi- some influence in the contrast perceived, we estimated the ratio
mum contrast representing highly visible points. Points were between the decoration area and the spider and evaluated the
manually tracked using the MTrackJ plugin in Fiji (Meijering relationship between the colour diversity and pattern complex-
et al. 2012). Subsequently, we used the X, Y coordinates from ity with the decoration ratio through simple linear regression.

13
23 Page 6 of 13 The Science of Nature (2023) 110:23

Fig. 2  Adult female of Argiope aurantia without a linear decora- column shows the results of the receptor noise limited method, per-
tion in its web. A Note that in this case the area of the web decora- forming pixel noise reduction after the acuity control based on Gauss-
tion is smaller compared with previous figure. B False colour image ian filters while preserving chromatic and luminance edges, simulat-
simulating the wasp Philanthus triangulum colour vision and created ing wasp spectral sensitivity (peaks are UV = 344, SW = 444, and
for visualisation purposes by assigning blue, green, and red for the MW = 524; Weber fraction = 0.13) and visual acuity (1.22 cpd) of
UV, SW, and MW wasp photoreceptors, respectively. C Results of the the wasp visual system at different distances; right column presents
quantitative colour pattern analysis (QCPA) of the spider as perceived the results of the achromatic edge intensity of the LEIA at different
by the wasp at different distances (10, 5, and 2 cm, respectively): left distances where different colours indicate different angle of contrast

Applying an ANOVA test, we compared the luminance and diversity (R2 = 0.26; F = 5.26; df = 11; P < 0.05). The abdo-
chromatic contrast of the abdomen and cephalothorax against men was more conspicuous than the cephalothorax against
the web, respectively, to know which part contribute the most to web decorations in terms of luminance contrast (F (1,24) =
the perceived contrast. The same method was applied to evalu- 7.089; P < 0.0001); Fig. 3A), but similar in chromaticity
ate chromatic and luminance kurtosis variation as descriptors contrast (F(1.24) = 1.637; P = 0.213). With respect to edge
of edge saliency at the different observation distances (2, 5, and saliency, luminance kurtosis declines with observation dis-
10 cm), and to estimate the differences between the body size tance (F(1,37) = 7.63; P < 0.05; Fig. 3B), but retains similar
perceived at three different positions during a cycle of web- chromatic saliency across all observation distances (F(1,37)
flexing behaviour. Finally, we compared the speed measured = 2.439; P = 0.127).
for the different tracked body parts using repeated measures The body parts tracked in the spider during one cycle of
ANOVA with pairwise paired t-tests using a Holm correction web-flexing behaviour (Fig. 4A) have a mean speed of 14.6
method as a post hoc test. All the statistical analyses were run cm/s (± 8.6 S.D.), but velocity differs between body parts
in R (R Development Core Team 2016). ­(F(3.96, 332.33) = 5.062; P < 0.0001). The abdomen posterior
(Abd P) and anterior (Abd A) ends moved faster (15.704
cm/s) than the legs III in both sides (LegIII L and LegIII
Results R) and the spider cephalothorax (head) (See Supplementary
Table 1).
Larger spiders had larger web decorations (R2 = 0.78; F Spider body size was measured as the area of a concave
= 43.66; df = 12; P < 0.001). Additionally, when viewed hull polygon during one cycle of web-flexing as perceived by
against natural backgrounds, higher decoration area/body an observer in front of the spider. We found that the perceived
size ratios correspond to more complex patterns (R2 = 0.443; area of the three spider reference positions during the cycle
F = 10.57; df = 11; P < 0.001; Figs. 1–2), and higher colour

13
The Science of Nature (2023) 110:23 Page 7 of 13 23

on the observer distance but also on web decoration ratio.

Additionally, while the spider’s abdomen exhibits the high-
est visual contrast when the spider is immobile, it is also the
body part that moves the most during web-flexing displays.
This heightened movement may startle predatory wasps and/
or make the spider look larger at key moments during their
deimatic display.
The interaction between Argiope spp. and araneopha-
gic wasps is not rare. Even though A. aurantia adults were
thought to be too big to be captured by most temperate North
American wasps (Blackledge and Wenzel 1999), juveniles
of A. aurantia and juveniles and adults of A. trifasciata were
preyed on by mud dauber wasps from the Trypoxylon spp.
genus (Muma and Jeffers 1945), and juveniles and adults of
A. argentata were found in nests of T. tridentatum (Robledo-
Ospina et al. 2021). Additionally, it has been reported that
the wasp Sceliphron caementarium hunts Argiope spiders
by bumping into the web, forcing the spider to drop. These
wasps then chase the spider by searching around in the
underlying vegetation in gradually enlarging circular pat-
terns (Blackledge and Pickett 2000). Because these spi-
ders exhibit high visual contrast against the background
Fig. 3  A Luminance contrast in just noticeable units (JND) of the (Robledo-Ospina et al. 2021), this suggests that visual cues,
body parts (i.e. abdomen and cephalothorax) of Argiope aurantia
against its web decoration as perceived by a wasp visual system using in conjunction with the chemical cues (Uma and Weiss
the receptor noise limited model (RNL) of Vorobyev and Osorio 2010), may be available for wasp predators seeking to prey
(1998). B Edge saliency through luminance contrast distribution (kur- on these spiders.
tosis) of Argiope aurantia in its web with decoration as perceived by The abdomen of Argiope aurantia is the body part with
a wasp visual system at different distances (2, 5, and 10 cm) using the
local edge intensity analysis (LEIA) according to Van der Berg et al. the highest contrast, and its conspicuousness is further
(2019). We used the mid wavelength (MW) photoreceptor for lumi- enhanced when viewed against web decorations. Similar to
nance perception in both cases. The whiskers represent the maximum our results, A. radon abdomen have high achromatic con-
and minimum values; the box, the interquartile range, and the notch trast against web decorations as seen by hymenopterans (Rao
represent the confidence interval of the median
et al. 2009). However, the saliency of the high-contrast pat-
tern edges reported here may function to disrupt body shape
were significantly different (Fig. 4B). The perceived spider recognition and detection (Cuthill et al. 2005, Schaefer and
size varies during the web-flexing behaviour (Fig. 4C), and Stobbe 2006, Stevens and Merilaita 2009). We suggest that
the spider was perceived as bigger in the front position (X2 = this may be the case for A. aurantia, at least for a hyme-
13.471, df = 2; P < 0.001; Fig. 4D). On the other hand, the nopteran visual system, because disruptive colouration like
perceived movement of the flexing behaviour by an observer this is thought to be effective in preventing improvements
in front of the spider is mainly composed of vertical vectors in camouflage breaking during search image formation by
in the optical flow (Fig. 5). Finally, when we included a third predators (Troscianko et al. 2018).
dimension depicting the web-flexing behaviour, we found evi- In the case of A. aurantia, in addition to a highly con-
dence that the spider position varies in depth (Fig. 6). spicuous abdomen, web decoration increases detectability
and contributes to a more visually complex signal even when
the spider remains still. This relation between the decoration
Discussion and the body colouration suggests that web decorations may
be critical to consider when investigating the antipredator
We evaluated the visual signals involved in the web-flexing visual signals of this species. The absence of web decora-
behaviour of Argiope aurantia as perceived by an approach- tions in nocturnal spiders supports a visually mediated func-
ing, visually oriented predatory wasp. We modelled the per- tion of these structures (Théry et al. 2011). Lubin (1974)
ception of the wasp, focused on chromatic and achromatic suggested that the infrequent occurrence of web decorations
vision, visual acuity, and a proxy of movement perception in Galapagos Island populations may be due to a lack of
(temporal resolution). Overall, we find that the general con- visually oriented predators. However, this should be inter-
trast and detectability of the spider varies depending not only preted cautiously because species do not rely only on one

13
23 Page 8 of 13 The Science of Nature (2023) 110:23

A C
450

400

Area mm^2
350

300

250

0 20 40 60 80
Time frames

B D
45
600

Area of Concave hull mm^2

40

500
35

400
30
Y

300
25

20
200

15 100
90 95 100 105 110 115 Back Rest Front
X

Position

Fig. 4  Body size changes during one cycle of web-flexing behaviour Changes in body size (area m ­ m2) in every frame recorded as might be
as might be perceived by an observer in front of the spider. A Loca- perceived by an observer in front the spider during one cycle of web-
tions tracked on the spider’s body during the web-flexing behaviour. flexing. D Perceived body size of the spider (polygon area ­mm2) at
B Spider body size represented by the concave hull polygons contain- the three positions of reference during one cycle of web-flexing. The
ing the body marks tracked in the spider body in the three positions whiskers represent the maximum and minimum values; the box, the
of reference in one cycle of web-flexing behaviour: rest position (grey interquartile range, and the notch represent the confidence interval of
polygon), back when the spider pushes backwards (red polygon), and the median
front when the spider moves towards the observer (blue polygon). C

secondary strategy. Some diurnal species possess a reper- might provide visual protection by breaking the body outline
toire of strategies, such as dropping to the ground, changing against web decoration (Wang et al. 2021b).
sides on the web, or fleeing from the web to a refuge. Bruce Many antipredator secondary strategies involve high con-
(2006), when writing about the controversy of the ecological trast fast-moving signals (e.g. deimatic displays) to deter
significance of web decoration, suggested that decorations the predator attack. Movement does not necessarily compro-
could potentially protect spiders by improving spider cam- mise camouflage; there are several ways in which prey may
ouflage, by making the spider look bigger to deter certain specifically incorporate motion into their defence strategy
gape-limited predators, or by acting as a physical barrier repertoire to compromise the cues used by their natural ene-
during predator attack. Nevertheless, we cannot ignore the mies (Tan and Elgar 2021). In addition to conspicuous high
body of evidence that suggests that web decoration functions contrast colouration, the fast movements of the A. aurantia
as a prey attraction device. Even though those hypotheses during web-flexing may present a challenge for predators in
are not mutually exclusive, solving this controversy will terms of image processing. In this regard, our results agree
require future research from a visual ecology perspective with several biological systems in which a deimatic display
(Eberhard 2020). involving high contrast colouration might be more effec-
With respect to the visual protection provided by dec- tive if the colouration is coupled to the defensive behaviour
orations in Argiope spp., available data suggests that the (Umbers et al. 2015, Umbers et al. 2017, Cox et al. 2021,
decoration of A. mascordi, even though conspicuous, could Martin et al. 2023). Here, we present evidence of a complex
provide some camouflage for spiders seen by hymenopter- scene in terms of pattern and colour diversity. When the
ans, either prey or predator (Bruce et al. 2005). Here, we spider flexes in the web, it might generate a sudden increase
suggest that in A. aurantia, disruptive high-contrast patterns in the visual stimuli input, particularly from the abdomen.
like those in the abdomen, which are difficult to visually In flying animals like wasps, visual motion detectors are
process and could therefore hamper body shape recognition, highly contrast-sensitive (Haag et al. 2004) and respond to

13
The Science of Nature (2023) 110:23 Page 9 of 13 23

compared to prey with background-matching patterns because

the stripes can reduce detection of moving prey by exploiting
the spatiotemporal limitations of predator perception.
The visual stimuli, here consisting of the spider, the deco-
ration, and the web-flexing behaviour, are the input to the
wasp visual system and are not static images. To compensate
for this dynamism in the visual scene, incessant eye and head
movements continuously reformat input signals, transform-
ing spatial patterns into temporal modulations on the retina
(Rucci et al. 2018). Understanding how movement perception
may drive animals’ behaviour is essential to determining the
function of spider web-flexing in the predatory responses by
wasps. For example, hymenopterans are known to use rela-
tive motion cues from the edges of objects to estimate their
height and guiding landing. Additional, compensatory head
roll movements while flying are driven by visual inputs based
on motion detection in some wasps (Viollet and Zeil 2013).
Because nearby objects are often more behaviourally rel-
evant than objects further away, object detection via optic
flow often weights objects according to their distance and,
thus, their functional relevance (Egelhaaf et al. 2012). Our
optical flow results suggest that when the spider engages
in web-flexing behaviour, the scene is mainly composed of
translational (vertical) vectors, which are highly relevant
for spatial information because they are distance-dependent
(Egelhaaf et al., 2012), and would exaggerate the percep-
tion of how close the spider is. This represents a dynamic
scene in which the spider’s general appearance and posi-
tion in the space changes as the predator approaches. More
specifically, web-flexing may make the spider appear to be
rapidly approaching the wasp, thereby inducing a looming
response and/or baffling the depth cues that a predator relies
on during attack.
There is evidence that sudden changes in body size
Fig. 5  Optical flow of Argiope aurantia while doing one cycle of web-
flexing as might be perceived by an observer in front of the spider. The
and colouration might deter predators from attacking via
arrows represent the vectors indicating the normalised magnitude (size deimatic displays (Martin et al. 2023). We propose that
and colour) and direction of the movement perceived the perceptual changes in the spider’s body size, speed,
and position, alongside the effect of approaching object,
may accentuate looming visual stimuli from the spider
spatiotemporally correlated changes in luminance (Derrington to the wasp and exploit innate behaviours based on basic
et al. 2004) that might occur from the moving edges of the neural processing to avoid collision of an “approaching”
target across the background and the internal contrast of the object. This could be explained by changes in the edges
target patterns (Umeton et al. 2019). In hymenopterans, the of the scene (spider + web decoration) when the loom-
edge detection is mainly mediated by the long-wavelength ing stimulus occurs (i.e. when the spider moves towards
receptors (MW) (Vasas et al. 2017). We used MW receptors the viewer looking bigger) that produces the ON (bright
for luminance perception, showing that the abdomen was the stimulation) and OFF (dark stimulation) antagonistic effect
most conspicuous body part, suggesting that this body region in the wasp’s motion perception because increasingly
may have evolved to interact with and/or confuse motion per- brighter-than-background shape increase neural response
ception by the MW receptors of its predators. The contrast of to approaching stimulus (Santer 2013), exploiting an
the background and body markings may affect the perception innate aversion to approaching objects.
of movement in several ways (Kelley & Kelley 2014). Ume- In conclusion, from a visual ecology perspective, a
ton et al. (2019) showed that prey with high-contrast stripes high-contrast abdomen in conjunction with rapid and sud-
become less visible when moving with sufficient speed when den movements towards (and away from) an approaching

13
23 Page 10 of 13 The Science of Nature (2023) 110:23

Fig. 6  3D tracking of the dif- Z Axis

A B
ferent body parts in Argiope
aurantia while doing one cycle 300

of the web-flexing behaviour.

A Marks tracked in the spider’s
Y Axis 200

body during the web-flexing

behaviour. B Trajectories 100

described by each mark tracked

in three dimensions. The col- 0

ours in the inset represent the

time sequence in frames during X Axis
one cycle of the web-flexing
behaviour. C Volume ­(cm3) of
C 250
the convex hull (see inset show-
ing a single leg, not to scale)
containing all the trajectories
200

Convex Hull Volume cm3

for each body part during one
cycle of web-flexing. The
convex hull is very flat due to 150
limited movement in the one of
the three dimensions
100

50

0
Leg2L Leg3L Leg4L Head AbdA AbdP Leg2R Leg3R Leg4R

flying predator can result in a highly visible yet dynamic Funding This project was supported by a Consejo Nacional de Ciencia
signal that threatens (Tolbert, 1975) and confuses wasp y Tecnología (CONACyT) Ciencia Básica grant (CB-2016-01/285529).
LR-O received support from CONACyT during his PhD program
predators (Caro and Ruxton 2019). Web-flexing behaviour (CONACyT-México 634812/338721).
might therefore have evolved to present a looming effect
that produces a recoil effect in wasp predators similar to Data availability statement The data will be archived and made avail-
responses documented in other predators (Shragai et al. able on GitHub (https://​github.​com/​dinrao/​Argio​pe).
S1 High-speed video (500 fps) showing the web-flexing behaviour
2017, Vagnoni et al. 2015). Furthermore, the web-flexing of Argiope aurantia. S2 high-speed video (500 fps) showing the marks
display may also interfere with the wasp’s optic flow per- tracked in the spider body to evaluate the kinematic behaviour. These
ception and flight manoeuvre control (Ache et al. 2019), are also available as supplementary materials online.
deterring the wasp from executing the final attack. Future
studies should evaluate this antipredator behaviour against Declarations
different flying insects in order to assess the potential costs Conflict of interest The authors declare no competing interests.
for the spider when performing the web-flexing behaviour.

Supplementary Information The online version contains supplemen-

tary material available at https://d​ oi.o​ rg/1​ 0.1​ 007/s​ 00114-0​ 23-0​ 1849-6.
References
Author contributions Luis E. Robledo-Ospina: conceptualization
(lead); data curation (lead); formal analysis (lead); investigation (lead); Ache JM, Polsky J, Alghailani S, Parekh R, Breads P, Peek MY, Bock
methodology (lead); project administration (equal); writing—original DD, Von Reyn CR, Card GM (2019) Neural basis for looming
draft (lead); writing—review and editing (equal). Nathan Morehouse: size and velocity encoding in the Drosophila giant fiber escape
conceptualization (supporting); methodology (supporting); writ- pathway. Current Biology 29(6):1073–1081.e4. https://​doi.​org/​
ing—review and editing (equal). Federico Escobar: conceptualization 10.​1016/j.​cub.​2019.​01.​079
(supporting); methodology (supporting); writing—review and editing Baird E, Kreiss E, Wcislo W, Warrant EJ, Dacke M (2011) Nocturnal
(supporting). Horacio Tapia-McClung: formal analysis (supporting); insects use optic flow for flight control. Biol Lett 7:499–501
methodology (equal); writing—review and editing (equal). Ajay Naren- Bateman PW, Fleming PA (2013) The influence of web silk decora-
dra: data curation (supporting); formal analysis (supporting); method- tions on fleeing behaviour of Florida orb weaver spiders, Argiope
ology (supporting); writing—review and editing (equal). Dinesh Rao: florida (Aranaeidae). Canadian Journal of Zoology 91:468–472
conceptualization (supporting); data curation (supporting); formal Blackledge TA, Pickett KM (2000) Predatory interactions between
analysis (supporting); methodology (equal); project administration mud-dauber wasps (Hymenoptera, Sphecidae) and Argiope
(equal); resources (lead); writing—review and editing (equal). (Araneae, Araneidae) in captivity. Journal of Arachnology
28:211–216

13
The Science of Nature (2023) 110:23 Page 11 of 13 23

Blackledge TA, Wenzel JW (1999) Do stabilimenta in orb webs attract Eisner T, Nowicki S (1983) Spider web protection through vis-
prey or defend spiders? Behavioral Ecology 10:372–376 ual advertisement: role of the Stabilimentum. Science
Bruce MJ (2006) Silk decorations: controversy and consensus. Journal 219:185–187
of Zoology 269:89–97 Endler JA (1991) Interactions between predators and prey. In: Krebs
Bruce MJ, Heiling AM, Herberstein ME (2005) Spider signals: are web JR, Davies NB (eds) Behavioural ecology: an evolutionary
decorations visible to birds and bees? Biology Letters 1:299–302 approach. Blackwell Scientific Publications, Oxford, pp 169–202
Caro T (2014) Antipredator deception in terrestrial vertebrates. Current Endler JA (2012) A framework for analysing colour pattern geom-
Zoology 60:16–25 etry: adjacent colours. Biological Journal of the Linnean Society
Caro T, Ruxton G (2019) Aposematism: unpacking the defences. 107(2):233–253
Trends in Ecology & Evolution 34:595–604 Endler JA, Cole GL, Kranz AM (2018) Boundary strength analysis:
Caro T, Argueta Y, Briolat ES, Bruggink J, Kasprowsky M, Lake J, combining colour pattern geometry and coloured patch visual
Mitchell MJ, Richardson S, How M (2019) Benefits of zebra properties for use in predicting behaviour and fitness. Methods
stripes: behaviour of tabanid flies around zebras and horses. in Ecology and Evolution 9:2334–2348
PLOS ONE 14:e0210831 Feller KD, Sharkey CR, Mcduffee-Altekruse A, Bracken-Grissom HD,
Caves EM, Johnsen S (2017) AcuityView: an R package for portray- Lord NP, Porter ML, Schweikert LE (2020) Surf and turf vision:
ing the effects of visual acuity on scenes observed by an animal. patterns and predictors of visual acuity in compound eye evolu-
Methods Ecol Evol 9:793–797 tion. Arthropod Structure & Development 60:101002
Cedhagen T, Björklund S (2007) Is web oscillation in the orb-web spi- Gawryszewski FM (2017) Anti-predator strategies. In: Viera C, Gon-
der Argiope lobata (Pallas, 1772) (Araneidae) an anti-predatory zaga MO (eds) Behaviour and Ecology of Spiders: Contributions
behaviour? Newsletter of the British Arachnological Society from the Neotropical region. Springer Nature, Switzerland, pp
109:10–11 397–415
Chakravarthi A, Rajus S, Kelber A, Dacke M, Baird E (2018) Dif- Gombin J, Vaidyanathan R, Agafonkin V (2017) concaveman: a very
ferences in spatial resolution and contrast sensitivity of flight fast 2D concave hull algorithm. R package version 1
control in the honeybees Apis cerana and Apis mellifera. Journal Gonzaga MO, Vasconcellos-Neto J (2005) Testing the functions of
of Experimental Biology 221:jeb184267 detritus stabilimenta in webs of Cyclosa fililineata and Cyclosa
Cheng R-C, Tso I-M (2007) Signaling by decorating webs: luring prey morretes (Araneae: Araneidae): do they attract prey or reduce the
or deterring predators? Behavioral Ecology 18:1085–1091 risk of predation? Ethology 111:479–491
Cloudsley-Thompson JL (1995) A review of the anti-predator devices Haag J, Denk W, Borst A (2004) Fly motion vision is based on Reich-
of spiders. Bulletin of the British Arachnological Society ardt detectors regardless of the signal-to-noise ratio. Proceedings
10:81–96 of the National Academy of Sciences of the United States of
Cox CL, Chung AK, Blackwell C, Davis MM, Gulsby M, Islam H, America 101:16333–16338
Miller N, Lambert C, Lewis O, Rector IV, Walsh M, Yamamoto Heathcote RJ, Troscianko J, Darden SK, Naisbett-Jones LC, Laker
AD, Davis Rabosky AR (2021) Tactile stimuli induce deimatic PR, Brown AM, Ramnarine IW, Walker J, Croft DP (2020) A
antipredator displays in ringneck snakes. Ethology 127:465–474 matador-like predator diversion strategy driven by conspicuous
Cronin TW, Johnsen S, Marshall J, Warrant EJ (2014) Visual ecology. coloration in guppies. Current Biology 30(14):2844–2851
Princeton University Press, Princeton, New Jersey Hoffmaster DK (1982) Predator avoidance behaviors of five species of
Cuthill IC, Stevens M, Sheppard J, Maddocks T, Párraga CA, Tros- Panamanian orb-weaving spiders (Araneae; Araneidae, Ulobori-
cianko TS (2005) Disruptive coloration and background pattern dae). Journal of Arachnology 10:69–73
matching. Nature 434:72–74 Horridge A (2019) The discovery of a visual system: the honeybee.
Defrize J, Théry M, Casas J (2010) Background colour matching by a CABI
crab spider in the field: a community sensory ecology perspec- How MJ, Zanker JM (2014) Motion camouflage induced by zebra
tive. Journal of Experimental Biology 213:1425–1435 stripes. Zoology 117:163–170
Derrington AM, Allen HA, Delicato LS (2004) Visual mechanisms Hughes AE, Troscianko J, Stevens M (2014) Motion dazzle and the
of motion analysis and motion perception. Annual Review of effects of target patterning on capture success. BMC Evolution-
Psychology 55:181–205 ary Biology 14:201
Donohue CG, Bagheri ZM, Partridge JC, Hemmi JM (2022) Fiddler Hughes AE, Jones C, Joshi K, Tolhurst DJ (2017) Diverted by dazzle:
crabs are unique in timing their escape responses based on speed- perceived movement direction is biased by target pattern orienta-
dependent visual cues. Current Biology 32:5159–5164.e4 tion. Proc R Soc B 284:20170015
Drinkwater E, Allen WL, Endler JA, Hanlon RT, Holmes G, Homziak Jackson RR (1992) Predator-prey interactions between web-invading
NT, Kang C, Leavell BC, Lehtonen J, Loeffler-Henry K, Ratcliffe jumping spiders and Argiope appensa (Araneae, Araneidae), a
JM, Rowe C, Ruxton GD, Sherratt TN, Skelhorn J, Skojec C, tropical orb-weaving spider. Journal of Zoology 228:509–520
Smart HR, White TE, Yack JE et al (2022) A synthesis of dei- Jackson RR, Rowe RJ, Wilcox RS (1993) Anti-predator defences of
matic behaviour. Biological Reviews 97(6):2237–2267 Argiope appensa (Araneae, Araneidae), a tropical orb-weaving
Eberhard W (1970) The predatory behavior of two wasps, Agen- spider. Journal of Zoology 229:121–132
oideus humilis (Pompilidae) and Sceliphron caementarium Kelley LA, Kelley JL (2014) Animal visual illusion and confusion:
(Sphecidae), on the orb weaving spider, Araneus cornutus the importance of a perceptual perspective. Behavioral Ecology
(Araneidae). Psyche 77:243–251 25:450–463
Eberhard WG (2020) Chapter 3. Function of orb web designs. In: Kodandaramaiah U, Palathingal S, Bindu Kurup G, Murali G (2020)
Eberhard WG (ed) Spider webs: Behaviour, function, and evo- What makes motion dazzle markings effective against predation?
lution. University of Chicago Press, Chicago Behavioral Ecology 31:43–53
Egelhaaf M (2023) Optic flow based spatial vision in insects. Journal Li D, Lee WS (2004) Predator-induced plasticity in web-building
of Comparative Physiology A 2023:1–21 behaviour. Anim Behav 67:309–318
Egelhaaf M, Boeddeker N, Kern R, Kurtz R, Lindemann JP (2012) Li D, Lim ML, Seah WK, Tay SL (2004) Prey attraction as a pos-
Spatial vision in insects is facilitated by shaping the dynamics sible function of discoid stabilimenta of juvenile orb-spinning
of visual input through behavioral action. Frontiers in Neural spiders. Animal Behaviour 68:629–635
Circuits 6:108

13
23 Page 12 of 13 The Science of Nature (2023) 110:23

Lubin YD (1974) Stabilimenta and barrier webs in the orb webs Schneider CA, Rasband WS, Eliceiri KW (2012) NIH Image to
of Argiope argentata (Araneae, Araneidae) on Daphne and ImageJ: 25 years of image analysis. Nature Methods 9:671–675
Santa Cruz Islands, Galapagos. The Journal of Arachnology Scott-Samuel NE, Baddeley R, Palmer CE, Cuthill IC (2011) Daz-
2:119–126 zle camouflage affects speed perception. PLOS ONE 6:e20233
Martin E, Steinmetz HL, Baek SY, Gilbert FR, Brandley NC (2023) Serres JR, Ruffier F (2017) Optic flow-based collision-free strategies:
Rapid shifts in visible Carolina Grasshopper (Dissosteira caro- from insects to robots. Arthropod Structure & Development
lina) coloration during flights. Front Ecol Evol 10: 900544 46:703–717
Mathis A, Mamidanna P, Cury KM, Abe T, Murthy VN, Mathis MW, Sheshadri S, Dann B, Hueser T, Scherberger H (2020) 3D recon-
Bethge M (2018) DeepLabCut: markerless pose estimation of struction toolbox for behavior tracked with multiple cameras.
user-defined body parts with deep learning. Nature Neuroscience Journal of Open Source Software 5:1849
21:1281–1289 Shragai T, Ping X, Arakaki C, Garlick D, Blumstein DT, Blaisdell
Meijering E, Dzyubachyk O, Smal I (2012) Methods for cell and par- AP (2017) Hermit crab response to a visual threat is sensitive
ticle tracking. Methods Enzymol 504:183–200 to looming cues. PeerJ 5:e4058
Muijres FT, Elzinga MJ, Melis JM, Dickinson MH (2014) Flies evade Soley FG (2019) A possible role of decorations in spiderwebs as
looming targets by executing rapid visually directed banked protection devices that distract predators. Rev Biol Trop
turns. Science 344:172–177 67:S164–S173
Muma MH, Jeffers WF (1945) Studies of the spider prey of several Spano L, Long SM, Jakob EM (2012) Secondary eyes mediate the
mud-dauber wasps. Annals of the Entomological Society of response to looming objects in jumping spiders ( Phidippus
America 38:245–255 audax, Salticidae). Biol Lett 8:949–951
Nakata K (2009) To be or not to be conspicuous: the effects of prey Srinivasan MV (1992) How bees exploit optic flow: behavioural
availability and predator risk on spider’s web decoration build- experiments and neural models. Philos Trans R Soc London
ing. Animal Behaviour 78:1255–1260 Ser B 337:253–259
Peitsch D, Fietz A, Hertel H, De Souza J, Ventura DF, Menzel R (1992) Stevens M, Merilaita S (2009) Defining disruptive coloration and
The spectral input systems of hymenopteran insects and their distinguishing its functions. Philos Trans R Soc London Ser
receptor-based colour vision. Journal of Comparative Physiol- B 364:481–488
ogy A 170:23–40 Stevens M, Merilaita S (2011) Animal camouflage: mechanisms and
Pekár S (2014) Comparative analysis of passive defences in spiders function. Cambridge University Press, Cambridge
(Araneae). Journal of Animal Ecology 83:779–790 Tan EJ, Elgar MA (2021) Motion: enhancing signals and concealing
Peron S, Gabbiani F (2009) Spike frequency adaptation mediates loom- cues. Biology Open 10:bio058762
ing stimulus selectivity in a collision-detecting neuron. Nat Neu- Temizer I, Joseph D, Baier H, Julia S (2015) A visual pathway for
rosci 12:318–326 looming-evoked escape in larval zebrafish. Current Biology
Peters RA, Clifford CW, Evans CS (2002) Measuring the structure of 25:1823–1834
dynamic visual signals. animal Behaviour 64:131–146 Théry M, Casas J (2009) The multiple disguises of spiders: web
R Development Core Team (2016) R: a language and environment for colour and decorations, body colour and movement. Philos
statistical computing. R Foundation for Statistical Computing Trans R Soc 369:471–480
Rao D, Cheng K, Herberstein ME (2008) Stingless bee response to Théry M, Debut M, Gomez D, Casas J (2005) Specific color sen-
spider webs is dependent on the context of encounter. Behav Ecol sitivities of prey and predator explain camouflage in different
Sociobiol 63:209–216 visual systems. Behavioral Ecology 16:25–29
Rao D, Webster M, Heiling AM, Bruce MJ, Herberstein ME (2009) Théry M, Insausti TC, Defrize J, Casas J (2011) The multiple dis-
The aggregating behaviour of Argiope radon, with special refer- guises of spiders. In: Stevens M, Merilaita S (eds) Animal
ence to web decorations. Journal of Ethology 27:35–42 camouflage: mechanisms and function. Cambridge University
Raudies F (2013) Optic flow. Scholarpedia 8:30724 Press, Cambridge, United Kingdom, pp 254–274
Robinson MH, Robinson B (1970) The stabilimentum of the orb web Tolbert W (1975) Predator avoidance behavior and web defensive
spider, Argiope argentata: an improbable defence against preda- structures in the orb weavers, Argiope aurantia and Argiope
tors. The Canadian Entomologist 102:641–655 trifasciata (Araneae, Araneidae). Psyche 82:29–52
Robledo-Ospina LE, Morehouse N, Escobar F, Falcón-Brindis A, Troscianko J, Stevens M (2015) Image calibration and analysis tool-
Jiménez ML, Rao D (2021) Prey colour biases of araneophagic box – a free software suite for objectively measuring reflec-
mud-daubing wasps. Animal Behaviour 172:25–33 tance, colour and pattern. Methods in Ecology and Evolution
Robledo-Ospina LE, Morehouse N, Escobar F, Rao D (2022) Search 6:1320–1331
image formation for spider prey in a mud dauber wasp. Behav- Troscianko J, Skelhorn J, Stevens M (2018) Camouflage strategies
ioural Processes 197:104619 interfere differently with observer search images. Proc R Soc
Rodríguez-Morales D, Tapia-Mcclung H, Robledo-Ospina LE, Rao D B 285:20181386
(2021) Colour and motion affect a dune wasp’s ability to detect Tyll S, Bonath B, Schoenfeld MA, Heinze H-J, Ohl FW, Noesselt T
its cryptic spider predators. Scientific Reports 11:1–9 (2013) Neural basis of multisensory looming signals. NeuroIm-
Rucci M, Ahissar E, Burr D (2018) Temporal coding of visual space. age 65:13–22
Trends in Cognitive Sciences 22:883–895 Uma DB, Weiss MR (2010) Chemical mediation of prey recognition
Ruxton GD, Allen WL, Sherratt TN, Speed MP (2018) Avoiding by spider-hunting wasps. Ethology 116:85–95
attack: the evolutionary ecology of crypsis, aposematism, and Umbers KD, Lehtonen J, Mappes J (2015) Deimatic displays. Current
mimicry. Oxford University Press, Oxford, United Kingdom Biology 25:R58–R59
Santer RD (2013) Motion dazzle: a locust’s eye view. Biol Lett Umbers KD, De Bona S, White TE, Lehtonen J, Mappes J, Endler
9:20130811 JA (2017) Deimatism: a neglected component of antipredator
Schaefer HM, Stobbe N (2006) Disruptive coloration provides cam- defence. Biol Lett 13:20160936
ouflage independent of background matching. Proc R Soc B Umeton D, Read JCA, Rowe C (2017) Unravelling the illusion of
273:2427–2432 flicker fusion. Biology Letters 13:20160831

13
The Science of Nature (2023) 110:23 Page 13 of 13 23

Umeton D, Tarawneh G, Fezza E, Read JCA, Rowe C (2019) Pat- Wang B, Yu L, Ma N, Zhang Z, Gong D, Liu R, Li D, Zhang S (2021a)
tern and speed interact to hide moving prey. Current Biology Conspicuous cruciform silk decorations deflect avian predator
29:3109–3113.e3 attacks. Integrative Zoology 17(5):689–703
Vagnoni E, Lourenco SF, Longo MR (2015) Threat modulates neural Wang B, Yu L, Ma N, Zhang Z, Liu Q, Fan W, Rong Y, Zhang S,
responses to looming visual stimuli. European Journal of Neu- Li D (2021b) Discoid decorations function to shield juvenile
roscience 42:2190–2202 Argiope spiders from avian predator attacks. Behavioral Ecol-
Valkonen JK, Vakkila A, Pesari S, Tuominen L, Mappes J (2020) Pro- ogy 32:1230–1239
tective coloration of European vipers throughout the predation Yilmaz M, Meister M (2013) Rapid innate defensive responses of mice
sequence. Animal Behaviour 164:99–104 to looming visual stimuli. Current Biology 23:2011–2015
Van Den Berg CP, Troscianko J, Endler JA, Marshall NJ, Cheney KL Zeil J (1993) Orientation flights of solitary wasps (Cerceris; Sphe-
(2019) Quantitative colour pattern analysis (QCPA): a compre- cidae; Hymenoptera). Journal of Comparative Physiology A
hensive framework for the analysis of colour patterns in nature. 172:207–222
bioRxiv 11:316–332 Zeil J (1997) The control of optic flow during learning flights. Journal
Vasas V, Hanley D, Kevan PG, Chittka L (2017) Multispectral images of Comparative Physiology A 180:25–37
of flowers reveal the adaptive significance of using long-wave-
length-sensitive receptors for edge detection in bees. J Comp Publisher’s note Springer Nature remains neutral with regard to
Physiol A 203(4):301-311 jurisdictional claims in published maps and institutional affiliations.
Viollet S, Zeil J (2013) Feed-forward and visual feedback control of
head roll orientation in wasps (Polistes humilis, Vespidae, Hyme- Springer Nature or its licensor (e.g. a society or other partner) holds
noptera). Journal of Experimental Biology 216:1280–1291 exclusive rights to this article under a publishing agreement with the
Vorobyev M, Osorio D (1998) Receptor noise as a determinant of col- author(s) or other rightsholder(s); author self-archiving of the accepted
our thresholds. Proc R Soc B 265:351–358 manuscript version of this article is solely governed by the terms of
Walter A (2019) Silk decorations in Argiope spiders: consolidation of such publishing agreement and applicable law.
pattern variation and specific signal function. Journal of Arach-
nology 47:271–275

13