Hemiptera is an ancient insect order (Permian) widely represented with nearly 100,000 species described in 145 families (Grimaldi, 2005; Gullan and Cranston, 2009). With the exception of a few predators and scavengers, Hemiptera are dominantly phytophagous (≈90% of species) and feed on a wide array of angiosperms and a few gymnosperms. They possess characteristic sucking mouthparts that they use to pierce plant tissues to access their feeding site (i.e., phloem, xylem or parenchyma) and a filtering chamber allowing them to extract liquid content (Carver et al., 1991). Psyllids (Psylloidea) belong to the monophyletic suborder Sternorrhyncha, a sister group of the rest of the Hemiptera, which includes aphids, whiteflies and scale insects. There are about 3000 species of psyllid worldwide (Hodkinson, 2009) of which approximately 350 are native to Australia (Hollis, 2004). The superfamily Psylloidea, which is more abundant and diverse in Australasia than in the northern hemisphere, is largely composed of acacia and eucalypt-feeding species, with the latter arguably having the closest associations with their host plants (Carver et al., 1991; Hollis, 2004; Hodkinson, 2009). Currently, the mechanisms psyllids employ to locate their host in complex mixed-vegetation habitats remain poorly understood.

Host selection by herbivorous insects revolves around successive steps starting from random foraging to location, recognition, assessment, and eventually acceptance or rejection of a plant as host for feeding and/or oviposition. All steps involve different senses and occur at varying distances from the plant surface. Since olfactory and visual cues are perceptible at distance, they play a determinant role in plant location by foraging insects (Schoonhoven et al., 1998). Plants produce a wide array of volatile compounds, which differ greatly in quality and quantity between species. These air-borne chemical “signatures” and their spatio-temporal variations offer a great variety of unique and specific cues that insects have evolved to recognize during their association with their host(s) (Visser, 1986; Bernays, 1998). In contrast, the diversity of visual cues reflected by plants is limited by the narrower variety of foliar pigments, which cause them to vary essentially in the “green-yellow-orange” part of the light spectrum visible to insects. As a result, olfaction is considered to provide more characteristic and reliable cues to insects in search of a specific host (Prokopy and Owens, 1983; Döring, 2014).

Nevertheless, most research indicates that vision is essential in aphid and whitefly orientation and most early studies failed to observe any chemotactic responses to host plant volatiles (Kennedy and Stroyan, 1959; Kennedy et al., 1961; Vaishampayan et al., 1975b; Pickett et al., 1992). More recent works have revealed behavioral and electrophysiological responses to host plant volatiles (Pickett et al., 1992; Visser et al., 1996; Pickett and Glinwood, 2007), which presumably assist host-alternating aphids during their migration and return to seasonal hosts. Similar observations were made with whiteflies (Vaishampayan et al., 1975b; Mellor and Anderson, 1995; Bleeker et al., 2009) although, as in aphids, the simultaneous application of visual stimuli was often required to elicit chemotactic responses to host plant odors.

One reason why many Sternorrhyncha rely to a lesser extent on olfaction probably resides in their poor abilities of active flight, which greatly hinders their capacity to follow airborne signals. Studies of flight behavior have shown that aphids and whiteflies use vision to control their groundspeed but can't sustain upwind flights in airflows exceeding 1 m.s⁻¹ (Kennedy and Stroyan, 1959; Kennedy and Thomas, 1974; Isaacs et al., 1999). Thus, their flight abilities are restricted in terms of distance (i.e., a few meters) and altitude. Migrations over long distances, observed during host alternation, are principally wind-assisted.

Visual perception of color stimuli has been shown to strongly influence aphids and whiteflies orientation. Most studies have shown that aphids (Kennedy et al., 1961; Moericke, 1969; Prokopy and Owens, 1983; Hardie, 1989; Döring, 2014) and whiteflies (Vaishampayan et al., 1975a,b; Coombe, 1982; Isaacs et al., 1999) orient preferentially toward green and yellow surfaces. Insect color perception is dependent upon the sensitivity of their photoreceptors. Since most insects possess the same chromophore (vitamin A1 based, 11-cis retinal), their sensitivity to light wavelength remains relatively conserved exhibiting maximal sensitivities in the UV, blue and green regions of the light spectrum (Briscoe and Chittka, 2001). Although it has not been as clearly demonstrated as with honeybees using psychophysics and electrophysiology (von Frisch, 1914; Daumer, 1956; Von Helversen, 1972; Peitsch et al., 1992), early studies demonstrated that aphids do exhibit intensity-independent behaviors in response to specific wavelengths (Moericke, 1955b) and are likely to possess a color-opponent mechanism (Moericke, 1950; Chittka and Döring, 2007) which are prerequisites for color vision (Menzel, 1979; Kelber et al., 2003). Moreover, recent works involving electrophysiological techniques have confirmed the existence of three photoreceptor types in at least two species of aphids (Kirchner et al., 2005; Döring et al., 2011). As a consequence, although it has not been convincingly demonstrated, aphids are suspected to have color vision (Döring and Chittka, 2007).

Unlike aphids and whiteflies, color preferences in psyllids have rarely been tested. This is mostly due to the fact that psyllids, like aphids, are presumed to be responsive to yellow-green (Mensah and Madden, 1992; Brennan and Weinbaum, 2001; Al-Jabr and Cranshaw, 2007; Hall et al., 2010), and therefore research has concentrated more on their olfactory responses (Valterovà et al., 1997; Soroker et al., 2004; Gross and Mekonen, 2005; Mayer et al., 2008; Patt and Sétamou, 2010). However, as in aphids, psyllids exhibit only weak chemotactic responses in the absence of visual cues (Hodkinson, 1974; Wenninger et al., 2009; Patt et al., 2011) and seem to possess an even less developed olfactory apparatus—which may only operate in host alternating species. An obvious limiting factor in the understanding of visual mechanisms in psyllids in psyllids is small size and the softness of their eyes, which has impeded the acquisition of physiological data.

In this study, we quantified the color preferences of four species of psyllid specialized on young to recently expanded leaves of different eucalypt hosts displaying distinct inter- and intra-specific optical characteristics. We sought to address the following topics: (1) whether psyllid color preferences are comparable to those exhibited by aphids and whiteflies; (2) whether color preferences are conserved between species; (3) whether the attraction to long wavelengths (red) exhibited by two of the species relies on stimuli intensity and/or their contrast against the background. Our ultimate objective, of which this study is an initial component, is to determine the ecological significance of vision in the utilization of eucalypts by psyllids. Red coloration of young leaves is not uncommon in eucalypts, which suggests that attraction to long wavelengths may occur in many other eucalypt-feeding psyllids.

This study is, to our knowledge, the first to comprehensively address psyllid responses to colors under controlled conditions and using stimuli of known reflectance. Ctenarytaina eucalypti and C. bipartita were attracted to “green” and “yellow” stimuli. Ctenarytaina bipartita preferred yellow and green significantly more than red but only green was significantly more attractive to C. eucalypti than red. The preference for yellow and green stimuli exhibited by these two species concurs with the responses of other Sterrnorhyncha such as aphids (Kennedy et al., 1961) and whiteflies (Vaishampayan et al., 1975b; Coombe, 1982). Moreover, similar species-specific differences in color preferences have been shown to occur in aphids (Moericke, 1969; A'brook, 1973; Kieckhefer et al., 1976) and curiously, according to season (Moericke, 1955a; Prokopy and Owens, 1983; Ramirez et al., 2008). Our results are consistent with field studies in which higher catches of psyllids were obtained using yellow and green sticky traps (Mensah and Madden, 1992; Brennan and Weinbaum, 2001; Al-Jabr and Cranshaw, 2007; Hall et al., 2010). The current findings also complement lab studies that have shown that olfactory responses to host plant volatiles are enhanced when experienced in conjunction with green or yellow stimuli (Wenninger et al., 2009; Patt et al., 2011).

By contrast, Anoeconeossa bundoorensis and Glycaspis brimblecombei strongly preferred “red” stimuli. “Yellow” was the second most chosen stimulus while “green” was only chosen occasionally. These observations contrast markedly with those of C. eucalypti and C. bipartita and more importantly with the literature pertaining to herbivorous insects. Since psyllid vision has received so little attention, it is currently impossible to estimate the incidence of attraction to long wavelength (red) stimuli among different species of psyllids, let alone more broadly within Hemiptera. High catches of the spruce aphid on red sticky traps is, to our knowledge, the only example of attraction to red within Sternorrhyncha (Straw et al., 2011). The preference of A. bundoorensis and G. brimblecombei for “red” forces us to consider that psyllid responses to color may not be limited to the usual yellow-green attraction observed in aphids.

The other question arising from our findings is how red is perceived by psyllids. Although most insects are known to have inherited the ancestral UV-blue-green set of photoreceptors (Chittka, 1996), their visual systems evolved different mechanisms to tune their spectral sensitivities to the cues most important to them. Thus, some insects, via the evolution of specific long wavelength photoreceptors and/or screening pigments, are able to perceive “red” as a true color. This is the case in numerous nectar-feeding lepidopterans and is not uncommon in other insect orders such as Odonata, Coleoptera and Hymenoptera. In most instances, the presence of “red” receptors is phylogenetically isolated and appears to be a consequence of convergent evolution (Briscoe and Chittka, 2001). Alternatively, “red” is perceived in the absence of specific red receptors via intensity-dependent stimulation of green photoreceptors. In the latter mechanism, called achromatic vision, “red” is perceived as a contrast or a level of brightness of the stimulus against its background. Achromatic insect vision has been shown to be involved in the detection of objects at long distance or under low light conditions. For instance, honeybees have been shown to rely alternately on chromatic and achromatic vision according to the visual angle of stimuli (Giurfa et al., 1997; Reisenman and Giurfa, 2008). Likewise, achromatic vision allows bees to detect red flowers although their conspicuousness is highly dependent on the background composition (Chittka and Waser, 1997). Some ant species have been shown to accurately discriminate between different color stimuli despite the apparent presence of a single photoreceptor type (Aksoy and Camlitepe, 2012). These examples demonstrate that attraction to a specific stimulus “color” may not be mediated by a color perception.

Given the current knowledge of hemipteran visual systems, psyllids have been presumed to be insensitive to long (red) wavelengths. However, our results show that, in the presence of a red and a dark gray (G45) stimulus of very low reflectance, two of the species of psyllids tested prefer “red” (Figure 5). This demonstrates that psyllids are somehow sensitive to long wavelengths, and such sensitivity promotes very clear preferences. Applying the extant wisdom, psyllids should not be able to discriminate between these two “dark” objects. One explanation for how insects might perceive long wavelength “red” radiation is via an achromatic mechanism as has been recently described in the rice leafhopper (Wakakuwa et al., 2014). This study revealed that under sufficiently high illumination conditions, leafhoppers could detect and orient toward long wavelengths stimuli when presented in a Y-maze apparatus. Electroretinograms and histological studies were conducted to prove that those responses were elicited in the absence of a specific long wavelength receptor, and that perception in the near infrared derived from the tail sensitivity of green receptors.

The possible role of a similar mechanism in psyllids was addressed in Experiment 3 by exposing psyllids to a variety of achromatic stimuli. Slight preferences for the darkest and brightest targets over the background indicate that psyllids used achromatic vision to locate the different gray stimuli. However, psyllids did not exhibit strong preference for any specific achromatic stimulus, in contrast to the very clear choices made in the presence of color stimuli. Calculations of intensity contrasts show that the dark gray (G45) offers a much greater contrast compared to the red stimulus (Figure 7A). Consequently, if contrast was the main factor influencing psyllid choice, dark gray (G45) would have been preferred over red. Likewise, reflectance calculations of the different stimuli reveal that the contrast of the “red” stimulus relative to the gray background is intermediate between that of G90 and G135 (Figures 7B,C), but neither of these targets was preferred. Hence, our results suggest that contrast of the stimuli against the background alone does not suffice for psyllids to prefer a specific stimulus. However, to appropriately address whether red can be perceived independently from intensity, responses of psyllids in the simultaneous presence of a red and various shade of grays need to be investigated.

Relative photon catches by putative UV-blue-green photoreceptors presented in Figure 8 show that, for all different sensitivity maxima tested, “red” stimulus associated stimulation varies only for the green photoreceptor. The intensity of stimulation of the green receptor would fluctuate between that of G45 (for λ_max = 530 nm) and that of G90 (λ_max = 560 nm). Consequently, under the assumptions of the UV-blue-green model, and attraction to red mediated by the achromatic pathway, psyllids would not be able to discriminate between red and G45 (Experiment 2) and would prefer either G45 or G90 (Experiment 3). Our modeling of potential achromatic mechanisms did not find strong support for an achromatic explanation. However, considering the paucity of physiological data in Sternorrhyncha for which spectral sensitivity of only three species is known to date (Mellor et al., 1997; Kirchner et al., 2005; Döring et al., 2011), the assumptions of the modeling of putative photoreceptors incite us to remain cautious. Whilst our current findings may be explicable by a trichromatic visual system consistent with the principles outlined in Wakakuwa et al. (2014), our results do suggest that it would be of value for future research to explore the possibility of a red-sensitive receptor in psyllid species using either electrophysiological or DNA-sequencing techniques.

Since the reflectance of leaves differs most noticeably in the green region of the spectrum and beyond (>500 nm), the addition of a red receptor to UV-blue-green receptors greatly increases the potential to detect different foliar hues (Lythgoe, 1979). For instance, red receptors in sawflies (Peitsch et al., 1992) and swallowtail butterflies (Kelber, 1999; Kelber et al., 2003; Kelber, 2006) have been shown to allow color-based discrimination of leaves of different age. Importantly, this capacity permits such insects to exhibit oviposition preferences that are based on leaf age. The existence of a red receptor in an individual species within a given taxon, e.g., in the oligolectic bee Callonychium petuniae (Wittmann et al., 1990), demonstrates that its occurrence can be sporadic and reflective of the relative influence of specific ecological interactions and such instances may explain why we observe statistically different behavioral differences between closely related psyllid species.

In summary, our results demonstrate that psyllids are sensitive and attracted to long wavelengths. Since our modeling and psyllid responses in the presence of achromatic stimuli do not allow us to exclude an intensity-independent attraction to long wavelengths, the mechanism mediating attraction to “red” remains to be resolved.

The ecological motives for “red” attraction are potentially significant. Red coloration in leaves is ontogenetically related, generally as an adaptive response to abiotic stressors (nutrient deficiency, drought, cold, photoinhibition and all of the aforementioned acting in conjunction with herbivory); it is produced by the synthesis of foliar anthocyanins. Since anthocyanins share the same biosynthetic pathway (i.e., phenyl propanoid pathway) as defensive compounds such as phenolics, their presence often coincides with chemically well defended leaves of low palatability (Lev-Yadun and Gould, 2009). The co-evolution theory proposes that red colors in autumnal reddening leaves evolved for aposematic ends to reduce herbivore loads (Hamilton and Brown, 2001; Archetti and Brown, 2004). But other theories posit the physiological benefits conferred by anthocyanins in reducing photo-oxidative stress during nutrient recovery from senescing leaves and the protection of photosynthetically active (young) leaves from excess light (Close et al., 2001; Schaefer and Wilkinson, 2004; Schaefer and Rolshausen, 2007). For deciduous plants, “red” advertises the imminence of leaf abscission. Consequently, the general consensus is that there is little imperative for herbivorous insect to be attracted to red leaves and typically anthocyanin-rich senescing leaves are not utilized by sucking insects (Kennedy et al., 1961; Chittka and Döring, 2007).

Although it has attracted much less attention, “red” coloration is also common in young, expanding leaves of many evergreen trees (Chalker-Scott, 2002; Karageorgou and Manetas, 2006; Manetas, 2006; Hughes, 2011) including those in the genus Eucalyptus (Sharma and Crowden, 1974; Thomas and Barber, 1974; Specht and Brouwer, 1975). Since variations in pigment concentrations are often associated with changes in leaf physiology, they should provide cues to psyllids and other insects reliant on visual location and assessment of hosts. Although examples are scarce, some studies have shown how insects can exploit visual cues to locate host leaves (Prokopy et al., 1983; Reeves, 2011). There is somewhat more evidence concerning how foliar pigments attract herbivores to leaves of a specific physiological state (Kennedy et al., 1961; Vaishampayan et al., 1975b; Prokopy and Owens, 1983; Shoonhoven et al., 2005). Hence, the discrimination of host from non-host plants based on visual cues is most likely to be the exception rather than the rule and probably depends importantly on the diversity and the architecture of the surrounding vegetation. However, in the context of small sucking insects with limited dispersal capacities, plant pigments could certainly act as cues by which to locate and orient toward of a preferred leaf type or age within a single host. Interestingly, the two red attracted psyllids are specialist-feeders on young foliage of the same host (Eucalyptus camaldulensis), which usually has red young and recently expanded leaves (see Figure 9). However, similar red foliage is not expressed by the hosts of the green-yellow attracted psyllids. Consequently, there already appears to be anecdotal evidence of tuning of psyllid visual responses mediated by their respective host's foliar colors. Thus, we believe that attraction to red could confer an advantage to psyllids searching for young leaves because it is not associated with leaf abscission. However, whether “red” host leaves elicit similarly strong attraction in natural conditions as artificial “red” color in our bioassay remains to be tested. Visual contrast dictated by color stimuli and the background upon which those are applied has been shown to influence aphid' attraction to “green” stimuli (Döring et al., 2004; Döring, 2014). Future studies should attempt to measure psyllid responses to more realistic stimuli under natural background conditions.

Following alighting on a plant, aphid host range is ultimately determined by gustatory cues which determine whether alates settle or resume their search for a host (Powell et al., 2006). Recent work has shown that nutrient availability and quality may also influence psyllid establishment on host leaves (Taylor, 1997; Steinbauer, 2013) suggesting a direct parallel with processes by which some aphids select yellow autumnal leaves with high concentrations of mobilized nitrogen (Holopainen and Peltonen, 2002; Chittka and Döring, 2007; Döring and Chittka, 2007). The associations between foliar reflectance spectra, the ontogeny of host eucalypt leaves and their nutritional quality are ongoing avenues of research by our group. Future papers of ours will address visually mediated host specificity in psyllids and whether foliar pigments provide information on plant physiology vital to their success.

Kevin Farnier, Adrian G. Dyer and Martin J. Steinbauer contributed to the experimental design. Experimental work and data analysis were performed by Kevin Farnier. Modeling work was done by Kevin Farnier with contribution from Adrian G. Dyer. Kevin Farnier wrote the paper which was revised by Martin J. Steinbauer and Adrian G. Dyer at early and final stages of the redaction.