Both authors contributed equally to this research.
The effect of colour variation in predators on the behaviour of pollinators: Australian crab spiders and native bees
Article first published online: 15 NOV 2010
© 2010 The Authors. Ecological Entomology © 2010 The Royal Entomological Society
Volume 36, Issue 1, pages 72–81, February 2011
How to Cite
LLANDRES, A. L., GAWRYSZEWSKI, F. M., HEILING, A. M. and HERBERSTEIN, M. E. (2011), The effect of colour variation in predators on the behaviour of pollinators: Australian crab spiders and native bees. Ecological Entomology, 36: 72–81. doi: 10.1111/j.1365-2311.2010.01246.x
- Issue published online: 7 JAN 2011
- Article first published online: 15 NOV 2010
- Accepted 4 October 2010First published online 15 November 2010
- Australian native bees;
- predator–prey coevolution;
- prey attraction;
- spider colour variation;
1. Australian crab spiders exploit the plant–pollinator mutualism by reflecting UV light that attracts pollinators to the flowers where they sit. However, spider UV reflection seems to vary broadly within and between individuals and species, and we are still lacking any comparative studies of prey and/or predator behaviour towards spider colour variation.
2. Here we looked at the natural variation in the coloration of two species of Australian crab spiders, Thomisus spectabilis and Diaea evanida, collected from the field. Furthermore, we examined how two species of native bees responded to variation in colour contrast generated by spiders sitting in flowers compared with vacant flowers. We used data from a bee choice experiment with D. evanida spiders and Trigona carbonaria bees and also published data on T. spectabilis spiders and Austroplebeia australis bees.
3. In the field both spider species were always achromatically (from a distance) undetectable but chromatically (at closer range) detectable for bees. Experimentally, we showed species-specific differences in bee behaviour towards particular spider colour variation: T. carbonaria bees did not show any preference for any colour contrasts generated by D. evanida spiders but A. australis bees were more likely to reject flowers with more contrasting T. spectabilis spiders.
4. Our study suggests that some of the spider colour variation that we encounter in the field may be partly explained by the spider's ability to adjust the reflectance properties of its colour relative to the behaviour of the species of prey available.
The signal communication between plants and their pollinators often occurs in such a way that both the transmitter and the receiver benefit from the signal produced. However, in several instances third-party organisms exploit this interaction. Despite the controversy about the effects of nectar robbers in plant–pollinator mutualism, they are a classical example of how a third organism exploits this interaction (Maloof & Inouye, 2000). Nectar robbers usually exploit plant–pollinator mutualism by piercing a hole in the corolla of the flower and drinking nectar without touching the pollen or stigma. Although the selective impact of nectar robbing on flower morphology has mostly been ignored (Maloof & Inouye, 2000), recent studies demonstrated that nectar robbers preferentially exploit flowers with long corolla tubes, reducing their reproductive success (Galen & Cuba, 2001; Castro et al., 2008; Navarro & Medel, 2009). Moreover, Ornelas et al. (2007) and Gomez et al. (2008) showed that there is a correlated evolution between nectar production and corolla tube length, which led some authors to suggest that the preference of nectar robbers for longer flowers could be explained by the higher nectar content of these flowers (Navarro & Medel, 2009).
There are further examples of plant–pollinator exploitation in spiders (Araneae), which have evolved several interesting forms of prey deception that lure prey with colours that resemble pollinator's food rewards. Under the right light conditions the orb-web spider Nephila clavipes produces webs with yellow silk that represents a stimulus for Trigona fluviventris bees searching for food: bees are more frequently attracted to yellow than to white webs under bright light and are less able to learn to avoid yellow webs (Craig et al., 1996). Furthermore, although there is still debate about the function of silk decorations in Argiope orb-web spiders, over the past decade several studies showed that these spiders also exploit the visual system of their prey by creating UV-reflective silk decorations that attract pollinators searching for food (Bruce et al., 2001, 2005; Li et al., 2004; Li, 2005; Cheng & Tso, 2007; Blamires et al., 2008; Tan et al., 2010). In addition, studies investigating the function of body colour markings in orb-web spiders demonstrated that spiders significantly reduced their foraging success when those colour markings were experimentally altered (Hauber, 2002; Tso et al., 2006, 2007; Chuang et al., 2007, 2008; Bush et al., 2008). All these recent findings support the idea that conspicuous body coloration in these spiders lure their prey.
In a similar way, several Australian crab spiders exploit the plant–pollinator mutualism by attracting and ambushing pollinators on flowers (Heiling et al., 2003; Heiling & Herberstein, 2004): they produce UV-reflective body colours that attract prey to the flowers they occupy. However, different species of pollinators react in a different fashion to Australian UV-contrasting spiders. European bees, Apis mellifera, approached and landed more on Chrysanthemum frutescens and Cosmos sp. flowers harbouring a Thomisus spectabilis and Diaea evanida spiders, respectively, compared with vacant flowers (Heiling et al., 2003; Herberstein et al., 2009). Similarly, the Australian native bee, Austroplebeia australis was also more likely to approach but less likely to land on C. frutescens flowers harbouring a T. spectabilis spider compared with vacant flowers (Heiling & Herberstein, 2004). These studies suggest that in the co-evolution between Australian native bees and crab spiders, the bees have evolved an anti-predatory response. The European honeybees, on the other hand, were introduced into Australia in 1822 (Hopkins, 1886) and feral colonies became widespread by 1860 (Laurie, 1886). Therefore honeybees have not had the opportunity to evolve a response to the deceptive Australian crab spider. To date, the response of native bees to the presence of native crab spiders has only been tested with one species of Australian bees and one species of Australian crab spider, which makes it difficult to extrapolate the suggestion that Australian bees have evolved an anti-predatory response towards crab spiders to native pollinators more generally.
It is difficult to interpret variation in pollinator response to crab spiders, because the colour signal produced by crab spiders is a plastic trait, and spiders change their colour over several days (Oxford & Gillespie, 1998). For example, Misumena vatia spiders turn from white to yellow in 10–25 days on artificially coloured backgrounds, and the reversed change can take 4–6 days (Gabritschevsky, 1927). Other studies have also reported similar colour changes for Misumenoides formosipes and Thomisus onustus crab spiders (Heckel, 1891; Gertsch, 1939). Australian T. spectabilis and D. evanida crab spiders can also change their body colour over several days between two colour morphs (UV-bright white morphs and UV-dull yellow morphs; A. L. Llandres, pers. obs.). Although spider UV reflection seems to vary broadly within and between individuals and species (Herberstein et al., 2009), we still do not know to what extent this variation occurs in natural populations of spiders. By quantifying colour variation in spider natural populations as well as the response of different species of pollinators to the individual variation in UV reflection of different species of crab spiders, we will be able to identify the ultimate mechanisms that maintain this colour trait in Australian crab spiders.
We currently know that different species of pollinators respond differently to the presence of a crab spider (Dukas & Morse, 2003, 2005; Robertson & Maguire, 2005; Gonçalves-Souza et al., 2008; Brechbühl et al., 2010). Therefore, including a community approach, in which crab spider's background colour matching is explored from the perspective of several main receivers in the field (community sensory ecology perspective), is necessary to understand this crab spider–flower visitor interactions (Brechbühl et al., 2010; Defrize et al., 2010). Furthermore, we also know that there is considerable variability in Australian crab spider coloration that creates high levels of variation in their visibility to prey and predators (Herberstein et al., 2009). Therefore, in order to disentangle these two factors, we looked at the natural variation in the coloration of Australian Thomisus spectabilis (Doleschall) and Diaea evanida (L. Koch) crab spiders and at the response of native pollinators to some of this variation.
Materials and methods
We measured the spectral reflectance (300–700 nm) of crab spiders and flowers using an optic fibre probe (Ocean Optics Inc., Dunedin, Florida) connected to a USB 2000 spectrometer (Ocean Optics Inc.). The USB 2000 spectrometer was connected to the PX-2 light source (Ocean Optics Inc.) and attached to a PC running OOIBase 32 spectrometer software (Ocean Optics Inc.).
We took five samples of each spider and flower and averaged them to calculate the photoreceptor excitation values (E) for the photoreceptors (UV, blue, and green) of honeybees (for methods see Chittka, 1992). We used honeybees (Apis melifera) as a visual model because the spectral sensitivity functions of Australian native bees are not known [note that for other bee species, such as bumblebees, the spectral sensitivities of the three receptor classes are very similar to the honeybees' receptors (Peitsch et al., 1992)]. The EUV, Eblue, and Egreen values describe the excitations by the UV, blue, and green photoreceptors and we used them to calculate the colour loci of spiders and their flower background in the bee colour hexagon. Then, we estimated the chromatic contrast between each pair of spider and flower by the Euclidean distance between the colour loci of the spider and the flowers in the colour hexagon (Chittka, 1992). Honeybees only use chromatic contrast for objects at short distances and they use the green photoreceptor (achromatic contrast) to discriminate an object from long distance (i.e. objects that subtend a small visual angle) (Giurfa et al., 1996; Spaethe et al., 2001). Hence, we also calculated achromatic contrast between honeybees and their background as the difference between the value of the green photoreceptor when excited by the spider and the value of the green photoreceptor when excited by the flower. Values greater than zero indicated that spiders were brighter than flower and values lower than zero indicated that flowers were brighter than spiders. In order to describe the excitation of UV and blue photoreceptors in the bees' retina we also calculate the specific contrast for these bee photoreceptors using the same method.
We collected female crab spiders Thomisus spectabilis (Thomisidae) (n = 79) from Bidens alba Linn. (Asteraceae) white daisies and female Diaea evanida (Thomisidae) (n = 95) sitting on Cosmos sp. (Asteraceae) yellow daisies. Thomisus spectabilis spiders and B. alba flowers were collected in the surrounding areas of Airlie Beach, Queensland (Australia), in April 2008 and D. evanida spiders and Cosmos sp. flowers were collected in suburban areas of Sydney, New South Wales (Australia), from February to March 2008. A total of eight flowers of each species were collected at random. We used a total of eight flowers of each species because data collected for the reflectance properties of each flower species, Cosmos sp. (N = 95) and B. alba (N = 111) flowers, have shown that the colour variation in these particular flower species is quite low (the mean ± SD values for the EUV, Eblue, and Egreen were 0.01 ± 0.01, 0.07 ± 0.03, and 0.70 ± 0.04 respectively for Cosmos sp. flowers and 0.42 ± 0.05, 0.80 ± 0.03, and 0.77 ± 0.03 respectively for Bidens alba flowers).
We calculated EUV, Eblue, and Egreen values of the flowers and spiders as viewed by the honeybee using the methodology described before (Chittka, 1992). For the eight flowers of each species of plants we measured the reflectance and averaged the eight E values of each plant species to obtain a natural mean background spectrum for each flower species. With the mean background spectrum we calculated the colour contrast and also the UV, blue, and green contrast to determine how each colour region contributes to the overall colour contrast created by the spider against a flower. Once caught, we maintained spiders in plastic containers with the flowers on which they were sitting in the field until we took their colour measurements within 5 days after spider collection. Previous experience has shown that this period in not enough to generate significant colour change in these spider species (A. L. Llandres, pers. obs.).
To compare the excitation values (E) of the honeybee UV, blue, and green photoreceptors between spiders and flowers collected in the field we used a t-test comparison for each of the three photoreceptors of honeybees and for each spider versus flower species comparison. We used a false discovery rate adjustment (α < 0.027) to account for the three non-independent calculations for the three photoreceptors of honeybees for each spider versus flower species comparison (Benjamini & Hochberg, 1995). We opted for the false discovery rate adjusted alpha instead of the Bonferroni adjustment because the Bonferroni adjustment has been shown repeatedly to be overly conservative (Benjamini et al., 2001; Narum, 2006).
Bee choice experiment
The bee choice experiment was carried out on the campus grounds of Macquarie University, Sydney, in March 2008. We used the Australian crab spider D. evanida, and Cosmos sp. flower species collected from the surrounding areas of Sydney in February 2008 and the Australian native bee Trigona carbonaria Smith (Apidae). The spiders used for the experiment were maintained in plastic containers in the laboratory against a constant dark background. They were fed with houseflies (Musca domestica) every week and watered daily. The native bees were maintained in an outdoor hive on campus and trained to visit a nectar feeder (30% sucrose solution), which consisted of a plastic jar (4 cm in diameter) placed upside down on a plastic lid. The days that the experiment was carried out the feeder was replaced by the experimental trials and between each trial the feeder was offered to the native bees again.
The experiment consisted of giving native bees a choice between two flowers, one of them occupied by a spider and the other without a spider. The spiders were anaesthetised with carbon dioxide and placed on a randomly selected flower of the pair. The spiders were placed on the petals in a way that resembled their natural hunting position. The flowers were placed on a black plastic lid and each pair of lids was positioned against a black background with a distance of 10 cm between the flower centres. The flower petals were cut to equalise the diameter of the flowers and their centre discs diameter were similar in size to ensure that the decision of the bees was not influenced by differences in flower traits. Each daisy and crab spider was used only once.
The experiment was carried out including olfactory cues (N = 37 choice trials) and excluding them (N = 35 choice trials). To exclude olfactory cues we covered the plastic lids where the flowers were placed with as see-through plastic foil (Glad Wrap®). The plastic foil is evenly permeable (<10% reduction) to all wavelengths of light between 300 and 700 nm (Heiling et al., 2003). We observed the number of native bees that approached the flowers within a distance of 4 cm during a period of 4 min and also observed the first bee that contacted one of the flower pair. After each day of the experiment we measured the spectral reflectance of spiders and flowers and calculated UV, blue, green, and colour contrast for each pair of spider and flower.
To confirm that the two flowers from each trial offered to native bees were similar in colour, we used a paired t-test to compare the EUV, Eblue, and Egreen of both petals and central disc between flowers with spider and vacant flowers. We further compared the number of bee approaches between spider-harbouring flowers and spider-free flowers using a matched paired t-test. Bee contact (first bee to contact the flower) was analysed with an independent binomial test. We used independent test for each comparison between flower with and without spider for the experiment with and without smell. We used a false discovery rate adjustment (α < 0.027) to account for the three non-independent calculations of the three photoreceptors of honeybees for flower petals and central discs comparisons between flowers with spider versus vacant flower.
Response of native bees to variation in spider contrasts
For these analyses, we used data from our bee choice experiment with D. evanida spiders and T. carbonaria native bees (hereafter Exp. 1) and also published data (Heiling & Herberstein, 2004) from an experiment with T. spectabilis spiders and Austroplebeia australis Friese (Apidae) native bees (hereafter Exp. 2). In these analyses we considered only the data of native bees approaching flowers when olfactory cues were included. We did this because there were too few landings for analysis and because native bees did not seem to respond well to the presence of the plastic foil in the experiment where odour was removed (see also Heiling & Herberstein, 2004). In order to test how native bees responded to the variation in colour contrast generated by the spiders against the flowers we used a regression analysis with the percentage of approaches to the flower with spider as the dependent variable. EUV, Eblue, Egreen and colour contrast generated by spiders against flowers were used as the independent variables.
Eight independent analyses were carried out for the data of the two bee choice experiments (four for each experiment). We performed an independent regression for all the independent variables (EUV, Eblue, Egreen and colour contrast) of each of the two experiments because all the independent variables were highly correlated with each other. We used a false discovery rate adjustment (α < 0.024) to account for the four regressions that we performed for each experiment.
To account for the non-normality of the residuals of our dependent variables we used Monte Carlo procedures to calculate empirical p-values of all the regressions (Davison & Hinkley, 2006). A total of 1999 Monte Carlo simulations were run using PopTools, version 3.0.6 (Hood, 2008) in Microsoft Excel 2007. Furthermore, we performed two independent partial least square (PLS) regression analysis (Carrascal et al., 2009), one for each of the two experiments, including percentage of approaches to the flower with spider as the dependent variable and the UV, blue, and green contrast as the independent variables. PLS analyses are especially useful when the predictor variables are highly correlated because this type of analysis allows us to include all the independent variables in a single PLS regression analysis to determine the weight with which each independent variable contribute to explain the dependent variable (for a detailed explanation see Carrascal et al., 2009).
The average values of the overall colour contrast created by T. spectabilis spiders against B. alba flowers and D. evanida spiders against Cosmos sp. flowers are above the value of 0.05 (Table 1 and Fig. 1), the theoretical detection threshold of honeybees (Thery & Casas, 2002). The overall colour contrast values are a product of individual values in the UV, blue, and green contrasts. Average UV, blue, and green contrasts for D. evanida spiders found on Cosmos sp. flowers and for T. spectabilis found on B. alba flowers are shown in Table 1.
|(a)||Spider (n = 95)||Flower (n = 8)||Contrast||t101||P|
|UV||0.13 ± 0.09||0.01 ± 0.00||0.12 ± 0.09||3.752||<0.001|
|Blue||0.38 ± 0.12||0.06 ± 0.01||0.32 ± 0.11||7.634||<0.001|
|Green||0.69 ± 0.05||0.71 ± 0.01||−0.02 ± 0.04||−1.041||0.300|
|Overall colour||—||—||0.31 ± 0.10||—||—|
|(b)||Spider (n = 79)||Flower (n = 8)||Contrast||t85||P|
|UV||0.65 ± 0.09||0.42 ± 0.02||0.23 ± 0.09||6.982||<0.001|
|Blue||0.82 ± 0.04||0.79 ± 0.01||0.03 ± 0.04||2.165||0.033|
|Green||0.79 ± 0.03||0.77 ± 0.01||0.02 ± 0.03||1.550||0.124|
|Overall colour||—||—||0.21 ± 0.07||—||—|
There was a significant difference in the excitation values between D. evanida spiders and Cosmos sp. flowers for the UV and blue photoreceptor (P < 0.001, Table 1, part a) but not for the green photoreceptor (P = 0.3, Table 1, part a). The difference in the photoreceptor excitation values between T. spectabilis spiders and B. alba flowers was only significant for the UV photoreceptor (P < 0.001; Table 1, part b).
Bee choice experiment
The excitation values of the flower petals and central disc between flowers with and without a spider did not show a significant difference on the UV, blue, or on the green photoreceptor excitation values for the flowers used for both experiments, with and without smell (Table 2).
|(a) With odour||n||Flower without spider (mean ± SD)||Flower with spider (mean ± SD)||t72||P|
|EUV||37||0.01 ± 0.01||0.01 ± 0.01||1.229||0.222|
|Eblue||37||0.08 ± 0.04||0.07 ± 0.04||1.287||0.202|
|Egreen||37||0.70 ± 0.01||0.70 ± 0.01||−1.182||0.241|
|EUV||37||0.01 ± 0.01||0.01 ± 0.01||0.604||0.547|
|Eblue||37||0.06 ± 0.04||0.06 ± 0.04||0.221||0.825|
|Egreen||37||0.53 ± 0.06||0.54 ± 0.07||−0.782||0.436|
|(b) Without odour||n||Flower without spider||Flower with spider||t68||P|
|(mean ± SD)||(mean ± SD)|
|EUV||35||0.01 ± 0.01||0.01 ± 0.01||0.987||0.326|
|Eblue||35||0.08 ± 0.04||0.07 ± 0.04||1.076||0.285|
|Egreen||35||0.70 ± 0.01||0.70 ± 0.02||−1.030||0.306|
|EUV||35||0.01 ± 0.01||0.00 ± 0.01||0.873||0.385|
|Eblue||35||0.07 ± 0.04||0.06 ± 0.05||0.564||0.574|
|Egreen||35||0.53 ± 0.07||0.54 ± 0.07||−0.312||0.755|
Despite the fact that both flowers (with and without spider) from each trial were similar in colour, our results show that in the experiment including olfactory cues native bees approached flowers with spider more often compared to vacant flowers (P = 0.033, Fig. 2). However, when the smell was excluded native bees approached flowers randomly (P = 0.523, Fig. 2). Native bees showed a non-significant tendency to contact vacant flowers more frequently than spider-harbouring flowers (P = 0.062 for experiment with olfactory cues and P = 0.055 for the experiment without olfactory cues, Fig. 2).
Response of native bees to variation in spider colour contrasts
The mean (±SD) UV, blue and green individual contrasts that spiders generated against flowers were 0.31 ± 0.22, 0.46 ± 0.18, and 0.03 ± 0.03 respectively for D. evanida spiders and Cosmos sp. flowers used for Exp. 1 and 0.15 ± 0.03, 0.01 ± 0.00, and 0.00 ± 0.00 respectively for T. spectabilis spiders and C. frutescens flowers used for Exp. 2.
The regressions showed non-significant relationships between percentages of T. carbonaria bee approaches and overall colour contrast as well as individual UV, blue, and green contrasts of D. evanida spiders (all P≥ 0.540, Table 3, part a). In contrast, in A. australis bees, the percentage of approaches to flowers with spider decreased significantly with individual UV contrast of T. spectabilis (P = 0.021 Fig. 3b). There were marginally significant negative relationships for the overall contrast and also for the individual blue and green contrast of the spiders (Table 3, part b and Fig. 3a,c,d).
|(a) Response of T. carbonaria bees towards D. evanida spider contrasts|
|(b) Response of A. australis bees towards T. spectabilis spider contrasts|
The PLS regression analyses showed that for the Exp. 1 all the components accounted for a very marginal proportion of the explained variance (all ≤5%, Table 4). However, for the Exp. 2 the PLS showed that the first component accounted for a major proportion of the explained variance (18%, Table 4), while the second and third components accounted for a marginal proportion of the explained variance (1% and 2% respectively, Table 4). The second and the third components work with the residual not explained by the first component, but, since their contribution was marginal, only the first component was considered to interpret the results. The meaning of the components can be interpreted considering the weights attained by each variable. The addition of the square of the weights within each component sums to one. Knowing this, if we focus on component one, the square weights with which each individual contrast contribute to explain the behaviour of A. australis bees were 0.37, 0.31, and 0.31 for the UV, blue, and green contrast respectively. This means that, for example, the UV contrast explained 37% of the total variance explained by the PLS regression model.
|W component 1||W component 2||W component 3|
Our study supports the idea that, unlike introduced pollinators (Heiling et al., 2003; Herberstein et al., 2009), Australian native bees are able to detect and avoid flowers harbouring crab spiders despite the fact that they are initially attracted to them. Our results showed that T. carbonaria native bees approached more but landed less on spider-harbouring flowers when odour was included. When we excluded olfactory cues native bees did not approach more frequently but were more likely to land on spider-free flowers. We suspect that the methods we used to exclude the odour may have affected native bee behaviour because bees seemed to be highly attracted to the plastic used to cover the flowers and they performed several inspection flights towards both flowers. Despite of that, we cannot rule out the possibility of an odour-based predator recognition mechanism in addition to the colour component. In fact, it is very likely that pollinating insects use more than one of these components to recognise and avoid their predators. Certainly, in a field study Reader et al. (2006) showed that A. mellifera bees responded to olfactory cues indicating the recent presence of a crab spider.
We found species specific differences in bee behaviour towards particular spider colour variation. Trigona carbonaria native bees did not show any preference for any of the individual colour contrasts generated by D. evanida spiders, but A. australis native bees showed a negative preference for flowers with more contrasting T. spectabilis spiders. It remains unclear why both bee species reacted differently to the extent of spider colour contrast, but we can think of three possibilities: firstly, there might be differences in the photoreceptors' spectral sensitivities between Apis mellifera and the two bee species used for this study; secondly, there may be differences in the visual system between T. carbonaria and A. australis bee species; and thirdly, the range of contrasts generated by both spider species used for the experiments was quite different and may have affected native bee behaviour. Further research is needed to confirm any interpretation of the differences in the behaviour of the two species of native bees presented in our study. According with our result, several studies reported that different species of prey reacted differently towards predatory cues (e.g. Sullivan et al., 2004; Lloyd et al., 2009). For example, Lloyd et al. (2009) showed that three species of skinks (Carlia rostralis, Carlia storri, and Carlia rubrigularis) reacted differently towards olfactory cues of a potential predator (Vanarus tristis goannas): although C. rostralis and C. storri skinks avoided the scent of the predator, C. rubrigularis did not show any avoidance behaviour towards predator olfactory cues.
In addition, our results showed that despite the negative preference for more UV-contrasting spiders, all the individual photoreceptor contrasts that T. spectabilis spiders generated against the flowers partly explained the behaviour of the A. australis native bees. Moreover, the PLS analysis showed that the three individual contrasts (UV, blue, and green) contributed almost equally to the response of A. australis behaviour towards T. spectabilis spiders. Other studies performed with the exotic bee A. mellifera have demonstrated that UV coloration in Australian crab spiders is the driving force that determines honeybee attraction towards T. spectabilis spiders (Heiling et al., 2003, 2005; Herberstein et al., 2009). In our study, however, we did not find that crab spider UV coloration was particularly important compared with other colours.
Taking into account our field data when examining how the overall colour contrast between spider and flower was created, both species presented a different pattern: in D. evanida spiders, differences in the UV and blue photoreceptor excitation values between flowers and spiders were crucial in generating overall contrast, whereas in T. spectabilis only the difference in UV photoreceptor excitation value contributed significantly to the overall colour contrast. In both spider species, the green photoreceptor excitation values were not different between spider and flower. Honeybees use the green photoreceptor (achromatic contrast) to discriminate objects from a long distance (i.e. objects that cover a small visual angle in the retina) (Giurfa et al., 1996; Spaethe et al., 2001). Our results indicate that, although Australian crab spiders are highly conspicuous when bees use their chromatic contrast to detect objects from a short distance, at a long distance they match the Egreen excitation of the flowers, which makes them highly camouflaged from their prey's perspective (Thery et al., 2005).
Considering our field data, the preference of Austroplebeia australis bees for low UV contrasting T. spectabilis spiders (Exp. 2) seems counterintuitive since most of the T. spectabilis spiders found in the field generated higher UV contrast values than those preferred by A. australis bees. This leads us to the following question: why are most spiders reflecting more UV in the field than the amount of UV that generates a more preferred contrast for certain species of native bees? We think that the answer to this question may lie in the availability of prey species. To date, it has only been suggested that variation in UV of Australian crab spider coloration could be the result of a trade-off between attracting prey and avoiding predators (Herberstein et al., 2009). Supporting this hypothesis, in a recent experiment Fan et al. (2009) found that the coloration pattern in the orb-web spider Nephila pilipes is the result of a trade-off between visually attracting prey and avoiding predators. In the case of crab spiders, because the more UV reflective spiders are, the more conspicuous they will be for both potential prey and predators (Heiling et al., 2005), high UV-reflective spiders may be more successful in the absence of predators than less UV-reflective spiders (Herberstein et al., 2009). However, our study highlights that some of this variation may indicate a colour strategy that matches the colour preferences or responses of the most abundant prey. Indeed, several studies have shown that different species of pollinators respond differently to the presence of a crab spider (Dukas & Morse, 2003, 2005; Robertson & Maguire, 2005; Gonçalves-Souza et al., 2008; Brechbühl et al., 2010). It, therefore, seems reasonable to assume that crab spiders present different strategies that increase foraging success according to the availability of prey species locally present. Accordingly, if the T. spectabilis spiders collected in the field for the present study were exposed mainly to Apis mellifera bees, it is parsimonious that most spiders generated a high UV contrast against the flowers, because high contrasting spiders would be more successful in attracting honeybees than low contrasting spiders (Heiling et al., 2005).
We propose that like other spiders, crab spiders may have evolved a foraging behaviour that exploits the colour cues that insects seek while searching for food (Craig & Bernard, 1990; Craig et al., 1996; Tso, 1996; Herberstein et al., 2000; Bruce et al., 2001, 2005). The variation in UV coloration of these spiders in different locations and at different times of the year might reflect the frequency of the most abundant prey and their species-specific colour response. Thus, the ability to up or down regulate UV would enable spiders to exploit whatever populations of prey are locally abundant. Some studies in animal body coloration showed that phenotypic plasticity in coloration allows animals to adjust their colour in response to specific types of predators (Hanlon et al., 1999; Templeton & Shriner, 2004; Stuart-Fox et al., 2006, 2008). For example, dwarf chameleons differently adjusted their coloration in response to two predator species that differed in their visual capabilities (Stuart-Fox et al., 2006, 2008). Following the same reasoning predators, and especially stationary predators, that are able to adjust their coloration to attract locally abundant species of prey are likely to increase their foraging performance. This, in turn, might help us to explain why different foraging strategies can be maintained in different populations of predators.
We believe that the selective advantage of exploiting different types of prey might have been one of the major forces influencing the evolution of UV coloration in Australian crab spiders and can explain the existing species specific variation in UV coloration as well as variation within different individuals of the same species. Moreover, our study highlights the importance of considering other colours than just UV to understand why crab spiders attract or deter certain species of prey. Our study also highlights the importance of studying background matching in the field from a community sensory ecology perspective (Defrize et al., 2010). Since each prey species has evolved specific visual abilities and behavioural responses to the same stimulus, by understanding crab spider background colour matching from the perspective of several main receivers in the field we will be able to better understand the function of background colour matching in these particular generalist predators.
We thank Anne Gaskett and Greg Holwell for technical support during the colour measurement and collection of spiders and Miguel Ángel Rodríguez-Gironés for statistical advice. This work was supported by the CSIC (studentship I3P-BPD2005 to ALL), Ministerio de Ciencia e Innovación/FEDER (project CGL2007-63223/BOS) and Macquarie University.
- 1995) Controlling the false discovery rate – a practical and powerful approach to multiple testing. Journal of the Royal Statistical Society Series B: Methodological, 57, 289–300. & (
- 2001) Controlling the false discovery rate in behavior genetics research. Behavioural Brain Research, 125, 279–284. , , , & (
- 2008) Why cross the web: decoration spectral properties and prey capture in an orb spider (Argiope keyserlingi) web. Biological Journal of the Linnean Society, 94, 221–229. , & (
- 2010) Ineffective crypsis in a crab spider: a prey community perspective. Proceedings of the Royal Society B: Biological Sciences, 277, 739–746. , & (
- 2001) Signalling conflict between prey and predator attraction. Journal of Evolutionary Biology, 14, 786–794. , & (
- 2005) Spider signals: are web decorations visible to birds and bees? Biology Letters, 1, 299–302. , & (
- 2008) Function of bright coloration in the wasp spider Argiope bruennichi (Araneae: Araneidae). Proceedings of the Royal Society B: Biological Sciences, 275, 1337–1342. , & (
- 2009) Partial least squares regression as an alternative to current regression methods used in ecology. Oikos, 118, 681–690. , & (
- 2008) Consequences of nectar robbing for the fitness of a threatened plant species. Plant Ecology, 199, 201–208. , & (
- 2007) Signaling by decorating webs: luring prey or deterring predators? Behavioral Ecology, 18, 1085–1091. & (
- 1992) The color hexagon – a chromaticity diagram based on photoreceptor excitations as a generalized representation of color opponency. Journal of Comparative Physiology A: Sensory Neural and Behavioral Physiology, 170, 533–543. (
- 2007) Diurnal and nocturnal prey luring of a colorful predator. Journal of Experimental Biology, 210, 3830–3837. , & (
- 2008) Deceptive color signaling in the night: a nocturnal predator attracts prey with visual lures. Behavioral Ecology, 19, 237–244. , & (
- 1990) Insect attraction to ultraviolet-reflecting spider webs and web decorations. Ecology, 71, 616–623. & (
- 1996) Evolution of predator–prey systems: spider foraging plasticity in response to the visual ecology of prey. American Naturalist, 147, 205–229. , & (
- 2006) Bootstrap Methods and their Applications, Cambridge Series in Statistical and Probabilistic Mathematics, 8th edn. Cambridge University Press, Cambridge, U.K. & (
- 2010) Background colour matching by a crab spider in the field: a community sensory ecology perspective. Journal of Experimental Biology, 213, 1425–1435. , & (
- 2003) Crab spiders affect flower visitation by bees. Oikos, 101, 157–163. & (
- 2005) Crab spiders show mixed effects on flower-visiting bees and no effect on plant fitness components. Ecoscience, 12, 244–247. & (
- 2009) Hunting efficiency and predation risk shapes the color-associated foraging traits of a predator. Behavioral Ecology, 20, 808–816. , & (
- 1927) Experiments on color changes and regeneration in the crab-spider Misumena vatia. Journal of Experimental Zoology, 47, 251–267. (
- 2001) Down the tube: pollinators, predators, and the evolution of flower shape in the alpine skypilot, Polemonium viscosum. Evolution, 55, 1963–1971. & (
- 1939) A revision of the typical crab spider (Misumeninae) of America north of Mexico. Bulletin of the American Museum of Natural History, 26, 277–442. (
- 1996) Detection of coloured stimuli by honeybees: minimum visual angles and receptor specific contrasts. Journal of Comparative Physiology A: Sensory Neural and Behavioral Physiology, 178, 699–709. , , & (
- 2008) Association between floral traits and rewards in Erysimum mediohispanicum (Brassicaceae). Annals of Botany, 101, 1413–1420. , , , , & (
- 2008) Trait-mediated effects on flowers: artificial spiders deceive pollinators and decrease plant fitness. Ecology, 89, 2407–2413. , , & (
- 1999) Crypsis, conspicuousness, mimicry and polyphenism as antipredator defences of foraging octopuses on Indo-Pacific coral reefs, with a method of quantifying crypsis from video tapes. Biological Journal of the Linnean Society, 66, 1–22. , & (
- 2002) Conspicuous coloration attracts prey to a stationary predator. Ecological Entomology, 27, 686–691. (
- 1891) Sur le mimétisme de Thomisus Onostus. Bulletin Scientifique de la France et de la Belgique, 23, 347–354. (
- 2004) Predator–prey coevolution: Australian native bees avoid their spider predators. Proceedings of the Royal Society of London Series B: Biological Sciences, 271, S196–S198. & (
- 2003) Crab-spiders manipulate flower signals. Nature, 421, 334. , & (
- 2005) The role of UV in crab spider signals: effects on perception by prey and predators. Journal of Experimental Biology, 208, 3925–3931. , , , & (
- 2000) The functional significance of silk decorations of orb-web spiders: a critical review of the empirical evidence. Biological Reviews, 75, 649–669. , , & (
- 2009) Evidence for UV-based sensory exploitation in Australian but not European crab spiders. Evolutionary Ecology, 23, 621–634. , & (
- 2008) PopTools, version 3.0.6 [WWW document]. URL http://www.cse.csiro.au/poptools. CSIRO, Canberra, Australia. (
- 1886) Illustrated Australasian Bee Manual. Auckland, New Zealand. (
- 1886) Landsborough's Exploration of Australia. Murby, London, U.K. (
- 2005) Spiders that decorate their webs at higher frequency intercept more prey and grow faster. Proceedings of the Royal Society B: Biological Sciences, 272, 1753–1757. (
- 2004) Prey attraction as a possible function of discoid stabilimenta of juvenile orb-spinning spiders. Animal Behaviour, 68, 629–635. , , & (
- 2009) Chemical discrimination among predators by lizards: responses of three skink species to the odours of high- and low-threat varanid predators. Austral Ecology, 34, 50–54. , & (
- 2000) Are nectar robbers cheaters or mutualists? Ecology, 81, 2651–2661. & (
- 2006) Beyond Bonferroni: less conservative analyses for conservation genetics. Conservation Genetics, 7, 783–787. (
- 2009) Relationship between floral tube length and nectar robbing in Duranta erecta L. (Verbenaceae). Biological Journal of the Linnean Society, 96, 392–398. & (
- 2007) Phylogenetic analysis of interspecific variation in nectar of hummingbird-visited plants. Journal of Evolutionary Biology, 20, 1904–1917. , , , & (
- 1998) Evolution and ecology of spider coloration. Annual Review of Entomology, 43, 619–643. & (
- 1992) The spectral input systems of hymenopteran insects and their receptor-based color-vision. Journal of Comparative Physiology A: Sensory Neural and Behavioral Physiology, 170, 23–40. , , , , & (
- 2006) The effects of predation risk from crab spiders on bee foraging behavior. Behavioral Ecology, 17, 933–939. , , & (
- 2005) Crab spiders deter insect visitations to slickspot peppergrass flowers. Oikos, 109, 577–582. & (
- 2001) Visual constraints in foraging bumblebees: flower size and color affect search time and flight behavior. Proceedings of the National Academy of Sciences of the United States of America, 98, 3898–3903. , & (
- 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. , & (
- 2008) Predator-specific camouflage in chameleons. Biology Letters, 4, 326–329. , & (
- 2004) Variation in the antipredator responses of three sympatric plethodontid salamanders to predator-diet cues. Herpetologica, 60, 401–408. , & (
- 2010) Why do orb-weaving spiders (Cyclosa ginnaga) decorate their webs with silk spirals and plant detritus? Animal Behaviour, 79, 179–186. , , , , , et al. (
- 2004) Multiple selection pressures influence Trinidadian guppy (Poecilia reticulata) antipredator behavior. Behavioral Ecology, 15, 673–678. & (
- 2002) Predator and prey views of spider camouflage – both hunter and hunted fail to notice crab-spiders blending with coloured petals. Nature, 415, 133. & (
- 2005) Specific color sensitivities of prey and predator explain camouflage in different visual systems. Behavioral Ecology, 16, 25–29. , , & (
- 1996) Stabilimentum of the garden spider Argiope trifasciata: a possible prey attractant. Animal Behaviour, 52, 183–191. (
- 2006) Function of being colorful in web spiders: attracting prey or camouflaging oneself? Behavioral Ecology, 17, 606–613. , , & (
- 2007) Nocturnal hunting of a brightly coloured sit-and-wait predator. Animal Behaviour, 74, 787–793. , & (