Colour Preference of the Deer Ked Lipoptena Fortisetosa (Diptera: Hippoboscidae)

Simple Summary Insects use visual stimuli to find habitats, food, or a mate while moving around. This trait might be exploited to intercept flying insects to monitor their populations and reduce their presence. Among the various visual stimuli, colours are commonly used to attract insects. Lipoptena fortisetosa is a hematophagous deer ectoparasite native to Japan that has spread to several central European countries and was recently recorded in Italy. Measures to monitor and control L. fortisetosa would be helpful given its potential threat as a pathogen vector for animals and humans. The objective of this research was to assess the potential use of colour to attract and trap L. fortisetosa. The response of the winged adults was evaluated through an experimental trial carried out in a wooded area of Tuscany using differently coloured sticky panels as traps. Blue panels attracted the highest number while yellow panels showed the lowest performance. This preference for blue could be useful in the design of traps to reduce the population of this parasitic fly which, at certain times, can reach a very high density, causing annoyance to wildlife and humans visiting natural areas. Abstract Lipoptena fortisetosa, a deer ked native to Japan, has established itself in several European countries and was recently recorded in Italy. This hippoboscid ectoparasite can develop high density populations, causing annoyance to animals and concern regarding the potential risk of transmitting pathogens to humans. No monitoring or control methods for L. fortisetosa have been applied or tested up to now. This research evaluated the possible response of L. fortisetosa winged adults to different colours as the basis for a monitoring and control strategy. In the summer of 2020, a series of six differently coloured sticky panels were randomly set as traps in a wooded area used by deer for resting. The results indicated a clear preference of the deer ked for the blue panels that caught the highest number of flies during the experimental period. Lower numbers of flies were trapped on the red, green, black, and white panels, with the yellow panels recording the fewest captures. The response clearly demonstrates that this species displays a colour preference, and that coloured traps might be useful for monitoring and limiting this biting ectoparasite in natural areas harbouring wildlife and visited by people.


Introduction
Lipoptena fortisetosa is a small blood-sucking ectoparasite whose primary hosts are various species of ruminant artiodactyl mammals, especially cervids and bovids [1], although it is also known to bite humans. Phylogenetically, it belongs to the subfamily Lipopteninae, but unlike other hippoboscids, it is not able to frequently switch host and attaches itself to a single animal for life. Newly emerged Lipoptena flies are fully winged adults that immediately search for a suitable host. Once found, the fly settles down to live in the host's fact, coloured traps coated with odourless glue are frequently used to sample different blood-feeding insects since they are inexpensive and easy to assemble [28].
The objective of this paper is to provide a fresh account on the response of L. fortisetosa to visual stimuli through an experiment carried out in a wooded area of Tuscany using differently coloured sticky panels as traps.

Materials and Methods
The field trial was conducted in a wooded area in Schignano (Prato, Tuscany, central Italy) at about 550 m a.s.l. (43.967432 N; 11.101761 E), where many warnings about the abundance of ked flies have been reported by people visiting this area. The study area consisted of a sloped clearing enclosed on three sides by a forest of mainly oak, holm oak, and chestnut that is frequently used by deer as a passageway or rest area.

Experimental Design
To evaluate the possible response of ked flies to colour, three series of differently coloured sticky traps were arranged in three different locations within the experimental area ( Figure 1). The first series bordered the forest to the southwest of the clearing and had an east-west orientation (trap sides with exposure north-south); the second was placed inside the woods and was oriented north-south (trap sides with east-west exposure); and the third series was placed to the north of the glade with the same orientation and trap exposure as the second series.
Insects 2021, 12, x 3 of 16 possible response to colours could be exploited to define monitoring and control strategies. In fact, coloured traps coated with odourless glue are frequently used to sample different blood-feeding insects since they are inexpensive and easy to assemble [28]. The objective of this paper is to provide a fresh account on the response of L. fortisetosa to visual stimuli through an experiment carried out in a wooded area of Tuscany using differently coloured sticky panels as traps.

Materials and Methods
The field trial was conducted in a wooded area in Schignano (Prato, Tuscany, central Italy) at about 550 m a.s.l. (43.967432 N; 11.101761 E), where many warnings about the abundance of ked flies have been reported by people visiting this area. The study area consisted of a sloped clearing enclosed on three sides by a forest of mainly oak, holm oak, and chestnut that is frequently used by deer as a passageway or rest area.

Experimental Design
To evaluate the possible response of ked flies to colour, three series of differently coloured sticky traps were arranged in three different locations within the experimental area ( Figure 1). The first series bordered the forest to the southwest of the clearing and had an east-west orientation (trap sides with exposure north-south); the second was placed inside the woods and was oriented north-south (trap sides with east-west exposure); and the third series was placed to the north of the glade with the same orientation and trap exposure as the second series. Each series consisted of three repetitions of six solid colours (black, transparent, blue, green, yellow, and red) in a randomized sequence for a total of 54 traps (18 traps per series) ( Figure 2). The colours were chosen on the basis of studies conducted on tsetse flies and other hematophagous insects [18,21,[29][30][31]. The transparent colour was used as a control. Coloured traps consisted of plastic alveolar polypropylene "plastonda" panels while transparent traps were made of Poliver (artificial glass polystyrene); all of them measured 20 cm × 30 cm × 2.5 mm (width, height, and thickness, respectively) and were purchased at the home improvement retailer OBI Italia. The spectral reflectance of these panels in the visible and UV regions, between 250 and 800 nm, was measured using a PerkinElmer Each series consisted of three repetitions of six solid colours (black, transparent, blue, green, yellow, and red) in a randomized sequence for a total of 54 traps (18 traps per series) ( Figure 2). The colours were chosen on the basis of studies conducted on tsetse flies and other hematophagous insects [18,21,[29][30][31]. The transparent colour was used as a control. Coloured traps consisted of plastic alveolar polypropylene "plastonda" panels while transparent traps were made of Poliver (artificial glass polystyrene); all of them measured 20 cm × 30 cm × 2.5 mm (width, height, and thickness, respectively) and were purchased at the home improvement retailer OBI Italia. The spectral reflectance of these panels in the visible and UV regions, between 250 and 800 nm, was measured using a PerkinElmer Lambda 1050 spectrophotometer coupled with a specific accessory for re-Insects 2021, 12, 845 4 of 16 flectance measurements (150 mm InGaAs Integrating Sphere) (Perkin Elmer Inc., Waltham, MA, USA). The black and transparent traps showed nearly constant reflectance over the wavelength range from 250 to 800 nm. The black colour exhibited a mean reflectance of about 6%, while for the transparent sample it was 17%. The blue trap displayed a maximum reflectance of 420-470 nm (∼60% of reflectance) while the green displayed 500-550 nm (∼20% of reflectance). The yellow panel showed a maximum reflectance of 520-550 nm (∼75% of reflectance), and the red, 620-650 nm (∼55% of reflectance) (Figure 3). Lambda 1050 spectrophotometer coupled with a specific accessory for reflectance measurements (150 mm InGaAs Integrating Sphere) (Perkin Elmer Inc., Waltham, MA, USA). The black and transparent traps showed nearly constant reflectance over the wavelength range from 250 to 800 nm. The black colour exhibited a mean reflectance of about 6%, while for the transparent sample it was 17%. The blue trap displayed a maximum reflectance of 420-470 nm (∼60% of reflectance) while the green displayed 500-550 nm (∼20% of reflectance). The yellow panel showed a maximum reflectance of 520-550 nm (∼75% of reflectance), and the red, 620-650 nm (∼55% of reflectance) ( Figure 3).   Lambda 1050 spectrophotometer coupled with a specific accessory for reflectance measurements (150 mm InGaAs Integrating Sphere) (Perkin Elmer Inc., Waltham, MA, USA). The black and transparent traps showed nearly constant reflectance over the wavelength range from 250 to 800 nm. The black colour exhibited a mean reflectance of about 6%, while for the transparent sample it was 17%. The blue trap displayed a maximum reflectance of 420-470 nm (∼60% of reflectance) while the green displayed 500-550 nm (∼20% of reflectance). The yellow panel showed a maximum reflectance of 520-550 nm (∼75% of reflectance), and the red, 620-650 nm (∼55% of reflectance) ( Figure 3).    Both sides of the panels were coated with a glue applied by brush (Planatol VP 1854 PSA, Ivog biotechnical systems GmbH, Neüsaß, Germany). In each series, traps were arranged 1.5 m above the ground and hung from two cords in a stationary position; the 18 panels within each series were placed close to each other. To avoid any interference from the juxtaposition of colours, the order was changed every two weeks according to a randomized experimental design.
The traps were set on 6 July 2020 and checked weekly: flies caught on each side were counted separately for every panel and then removed to estimate their response to colours. The traps remained continuously exposed until 29 October 2020, and this time period was chosen based on L. fortisetosa adults in Europe being reported as present from June to October [5].
On the same day as the trap control, sweeping paths were performed in different environments near the experimental area (woodland, forest edge, open field, track).

Statistical Analyses
Analyses were carried out considering the mean number of flies caught by each trap per day as a dependent variable and the following as independent variables: the position of the series (position 1, position 2, and position 3), trap colour, sampling date, and trap activity (dichotomised as traps that captured L. fortisetosa or did not). From 1 October onwards, most of the traps were inactive and an extremely low number of flies (9) was caught. For this reason, only data from 15 July to 1 October were analysed.
To highlight potential differences in trap attractiveness, the data structure was checked through factor analysis of mixed data (FAMD), and then inferential statistics were applied.
The goal of the FAMD analysis was to explore the association between all the variables and highlight which factors determined the variability of the average number of flies caught by the traps. FAMD was performed with the open-source software RStudio (RStudio Version 1.3.1093, 2009-2020, PBC, Boston, MA; http://www.rstudio.com/) using the packages FactoMineR and FactoExtra. All variables were considered active, and missing data were managed using the package missMDA [32,33].
After the FAMD observation, inferential analysis was carried out to highlight differences among active traps compared to inactive traps as well as the average number of caught flies.
The proportion of active with respect to inactive traps was analysed using a Chisquare test (6 × 3 contingency tables) for the null hypothesis that all proportions were equal. When necessary, multiple pairwise comparisons were applied to the number of active traps according to the variable series position (H 0 : p 1 = p 2 ; H 0 : p 1 = p 3 ; and H 0 : p 2 = p 3 ) and to the variable trap colour (H 0 : p 1 = p 2 = . . . = p 6 ). The Type I error rate was adjusted, thereby reducing the maximum error rate of 0.05 by the total number of comparisons [34]. Chi-square tests were carried out using Excel software (Microsoft 365, 2016, Microsoft Italia, Milano, Italy).
Differences in the average number of L. fortisetosa captured on traps of different colours were analysed using univariate analysis of variance with the sampling dates from 15 July to 1 October and panel colours as factors, while the dependent variable was the average number of L. fortisetosa captured per day. Data were log(x + 1) transformed [34], and a pairwise comparison was performed using the Bonferroni test if a main effect was highlighted. All statistical analyses were performed using IBM SPSS Statistics for Windows, Version 25.0 (IBM Corp, Armonk, NY, USA) software.

Overall Evaluation of the Experiment Set on Lipoptena fortisetosa Captures
The eigenvalue to explain the total variance of data was selected according to the criteria defined by Karlis et al., [35]. Dimension 1 (Dim. 1) satisfied the criterion (eigenvalue >1.31) and summarized most information; however, to interpret the FAMD results, the second dimension (Dim. 2) was also considered. Both dimensions explained 17.04% of the overall variability. Although it was less than 50%, the representation was valid for describing the variance distribution within the dataset. The results of the FAMD analysis are reported in Table 1.   Table 1).
icantly higher number of active traps in position 1 (71.48%) compared to position 2 (37.50%) (χ 2 = 20.88, df=5, p < 0.001) and position 3 (20.37%) (χ 2 = 15.85, df = 5, p = 0.007). The number of active traps in position 3, where the lowest values were recorded, was comparable to the number of active traps in position 2 (χ 2 = 9.74, df = 5, p = 0.08). Considering the variable colour, there was a similar proportion of active compared to inactive traps (χ 2 = 11.23, df = 5, p = 0.34) ( Figure 5).   The FAMD results emphasized that position 1 was more appropriate for insect monitoring since it explained most of the variability and was linked to active traps. Moreover, 12 and 20 August seemed to be the most favourable days for L. fortisetosa capture. The outcomes suggest that there were optimal capturing periods that should be taken into account. The traps with the most divergent results were the blue and the yellow ones, with the blue traps being the most active.

Colour Preference of Lipoptena fortisetosa Winged Adults.
The FAMD and trap activity results show low effectiveness for series positions 2 and 3; thus, only L. fortisetosa captured by traps in position 1 were analysed. The log-transformed data of the captured L. fortisetosa highlighted significant differences among trap colours (main effect colours F = 15.82, p < 0.001) and sampling dates (main effect of the sampling dates F = 43.54, p < 0.001). Differences among the average number of flies caught daily by the panels were consistent across all sampling periods (interaction effect F = 1.42, p = 0.06). The blue panels recorded the highest number of individuals with an average daily catch of 1.28 ± 1.08 (mean ± SE). On the other hand, the yellow traps showed the lowest average values (0.33 ± 0.29 mean ± SE). The blue and yellow colours were significantly different from all the other panel colours (black, green, red, and transparent), which showed a similar average number of insects (respectively: 0.63 ± 0.58, 0.77 ± 0.72, 0.80 ± 0.88, and 0.64 ± 0.10 mean ± SE) ( Figure 6).

Colour Preference of Lipoptena fortisetosa Winged Adults
The FAMD and trap activity results show low effectiveness for series positions 2 and 3; thus, only L. fortisetosa captured by traps in position 1 were analysed. The log-transformed data of the captured L. fortisetosa highlighted significant differences among trap colours (main effect colours F = 15.82, p < 0.001) and sampling dates (main effect of the sampling dates F = 43.54, p < 0.001). Differences among the average number of flies caught daily by the panels were consistent across all sampling periods (interaction effect F = 1.42, p = 0.06). The blue panels recorded the highest number of individuals with an average daily catch of 1.28 ± 1.08 (mean ± SE). On the other hand, the yellow traps showed the lowest average values (0.33 ± 0.29 mean ± SE). The blue and yellow colours were significantly different from all the other panel colours (black, green, red, and transparent), which showed a similar average number of insects (respectively: 0.63 ± 0.58, 0.77 ± 0.72, 0.80 ± 0.88, and 0.64 ± 0.10 mean ± SE) ( Figure 6).
Insects 2021, 12, x 9 of 16 0.33 ± 0.06, and 0.34 ± 0.06) and after 26 August. On other days, the averages were statistically similar; however, 9 September (0.68 ± 0.12) was an exception, with a slight increase in the daily catch. This value was similar to those recorded on 6 and 26 August (Figure 7). In Figure 8, which shows the trend of flies trapped by the blue panels over the whole period, the average number of L. fortisetosa was considerably relevant, reaching the highest value of 22 specimens per trap. Moreover, after the August peak, the blue traps also caught flies effectively in September although that peak was less than half (8 insects/trap).  Regarding the variable sampling date, the most significant numbers of L. fortisetosa were recorded in August (mean ± SE of 1.09 ± 0.16, 1.49 ± 0.21, 1.90 ± 0.21, and 0.82 ± 0.14 respectively for 6, 12, 20, and 26 August). The averages recorded on these days were significantly higher than the values obtained on earlier dates (15-27 July: avg ± SE 0.19 ± 0.03, 0.33 ± 0.06, and 0.34 ± 0.06) and after 26 August. On other days, the averages were statistically similar; however, 9 September (0.68 ± 0.12) was an exception, with a slight increase in the daily catch. This value was similar to those recorded on 6 and 26 August (Figure 7). Insects 2021, 12, x 9 of 16 0.33 ± 0.06, and 0.34 ± 0.06) and after 26 August. On other days, the averages were statistically similar; however, 9 September (0.68 ± 0.12) was an exception, with a slight increase in the daily catch. This value was similar to those recorded on 6 and 26 August (Figure 7). In Figure 8, which shows the trend of flies trapped by the blue panels over the whole period, the average number of L. fortisetosa was considerably relevant, reaching the highest value of 22 specimens per trap. Moreover, after the August peak, the blue traps also caught flies effectively in September although that peak was less than half (8 insects/trap).  In Figure 8, which shows the trend of flies trapped by the blue panels over the whole period, the average number of L. fortisetosa was considerably relevant, reaching the highest value of 22 specimens per trap. Moreover, after the August peak, the blue traps also caught flies effectively in September although that peak was less than half (8 insects/trap).

Colour Preference
Colours tested in this study can be qualified or distinguished based on their spectral characteristics: yellow and red fit into the so-called "cut-off" colours, with a steeply sloped spectrum; blue and green relate to the "band reflecting" colours with discrete reflectance peaks; transparent and black, having constant reflectance at any wavelength, can be considered neutral [19,36].
As far as we know, most dipterans possess five types of photoreceptors, which can be sensitive in a wide band of wavelengths from UV up to the green wavelength range [37,38], so we speculated that L. fortisetosa followed this general model, although the spectral sensitivity of the deer keds has never been measured.
The capture of L. fortisetosa in Schignano (Prato, Italy) using differently coloured traps clearly demonstrates that this parasitic species displays a colour preference, since the winged adults selected colours according to a preference scale. Blue traps caught a significantly higher number of adults, compared to black, red, green, and transparent traps, while yellow traps seemed to be the least attractive. This colour ranking was consistent throughout the experimental period. Colour attractiveness has, to date, been poorly investigated in hippoboscid flies, with the only reported observations on L. cervi in Finland [21] where black-coloured clothing worn by moving people was compared to white, blue, and red clothing by evaluating the number of flies that landed on them or were caught by transparent flypaper. In the field trials, deer keds preferred black to white clothing, but showed no significant preferences for blue or red.
Our field trial results on L. fortisetosa differ from those of Kortet et al. [21] in that a blue colour preference was determined, though there are differences in the experimental design (coloured plastic panels in a stationary position instead of coloured clothing worn by moving people) and Lipoptena species. In the experiment carried out in Finland, other factors influencing host location may have affected deer ked choice, such as movement, chemical cues, and host size. Usually, moving objects are detected by an achromatic photoreceptor channel [39]; however, many insects are attracted by moving objects with particular colours showing the so-called "wavelength-specific behaviour" [40]. Different long-distance stimuli could also have influenced these flies, such as carbon dioxide, or odour cues emitted by the host skin as was demonstrated for different groups of hematophagous dipterans [13]. However, it is difficult to separate the effect of different stimuli

Colour Preference
Colours tested in this study can be qualified or distinguished based on their spectral characteristics: yellow and red fit into the so-called "cut-off" colours, with a steeply sloped spectrum; blue and green relate to the "band reflecting" colours with discrete reflectance peaks; transparent and black, having constant reflectance at any wavelength, can be considered neutral [19,36].
As far as we know, most dipterans possess five types of photoreceptors, which can be sensitive in a wide band of wavelengths from UV up to the green wavelength range [37,38], so we speculated that L. fortisetosa followed this general model, although the spectral sensitivity of the deer keds has never been measured.
The capture of L. fortisetosa in Schignano (Prato, Italy) using differently coloured traps clearly demonstrates that this parasitic species displays a colour preference, since the winged adults selected colours according to a preference scale. Blue traps caught a significantly higher number of adults, compared to black, red, green, and transparent traps, while yellow traps seemed to be the least attractive. This colour ranking was consistent throughout the experimental period. Colour attractiveness has, to date, been poorly investigated in hippoboscid flies, with the only reported observations on L. cervi in Finland [21] where black-coloured clothing worn by moving people was compared to white, blue, and red clothing by evaluating the number of flies that landed on them or were caught by transparent flypaper. In the field trials, deer keds preferred black to white clothing, but showed no significant preferences for blue or red.
Our field trial results on L. fortisetosa differ from those of Kortet et al. [21] in that a blue colour preference was determined, though there are differences in the experimental design (coloured plastic panels in a stationary position instead of coloured clothing worn by moving people) and Lipoptena species. In the experiment carried out in Finland, other factors influencing host location may have affected deer ked choice, such as movement, chemical cues, and host size. Usually, moving objects are detected by an achromatic photoreceptor channel [39]; however, many insects are attracted by moving objects with particular colours showing the so-called "wavelength-specific behaviour" [40]. Different long-distance stimuli could also have influenced these flies, such as carbon dioxide, or odour cues emitted by the host skin as was demonstrated for different groups of hematophagous dipterans [13]. However, it is difficult to separate the effect of different stimuli and to ascribe, with certainty, the location of a host to movement per se, since a moving subject may also release more carbon dioxide or may be warmer than one that is stationary [41].
In other hematophagous Diptera, such as Glossinidae (tsetse flies), blue proved to be the most attractive colour over a short distance [36]. Tsetse and ked flies show quite similar mouthparts and feeding activity [4,42], but once a suitable host is found they display different parasitic behaviour. Unlike deer keds, Glossina species frequently change host.
Additionally, the morphological affinity between these families could also pertain to eye structure and colour preference, which have been thoroughly investigated in tsetse flies [43,44]. In a field trial in Zimbabwe, Glossina morsitans morsitans and G. pallidipes showed a marked preference for traps covered by royal blue cotton, strongly reflecting blue-green wavelength bands, but poorly reflecting ultraviolet or green-yellow-orange bands [19]. A similar behaviour was described for another important tsetse fly, Glossina fuscipes fuscipes, which was particularly attracted by blue cloth panels, especially phthalogen blue panels, while yellow was the least attractive [36]. However, olfactory stimuli can affect the attraction of tsetse flies to different colours [17].
Other flies of economic importance, such as Stomoxys calcitrans, were investigated to highlight possible colour preferences that could be used to improve monitoring and control interventions. Experiments with blue and black targets with different patterns demonstrated the importance of blue in attracting flies of this species [45]. In addition, the attraction of S. calcitrans to polyethylene blue screens was not affected when different hues of blue were used [31]. Phthalogen blue with spectral sensitivity at 350, 450, and 625 nm markedly influenced Stomoxys spp. capture [31,46].
Other blood-sucking dipterans have shown a response to colours partly comparable to that exhibited by L. fortisetosa. For example, in Canada, Tabanidae and Simuliidae were tested with differently shaped and coloured traps and generally responded more to colours than to shapes [29]. Remarkably, all the investigated species belonging to both these families were consistently not attracted by yellow silhouettes. Moreover, when subjected to solid blue and solid yellow traps, tabanids were more attracted by the blue ones, and when subjected to blue/yellow striped traps, they always chose the blue part. Similar outcomes were highlighted in other studies [30] all of these tabanid data were quite consistent with our observations on L. fortisetosa, especially for the poor attractiveness of yellow. Nevertheless, for some herbivorous dipterans, such as Tephritidae and especially the Mexican fruit fly, Anastrepha ludens, laboratory trials showed a particular preference for yellow together with green, over black, red, blue, and white [47]. Similarly, yellow traps are commonly used in monitoring and control strategies in different olive growing areas for the olive fruit fly, Bactrocera oleae [48,49]. The recognition of green fruit by fruit flies, as well as the detection of green leaves by other herbivorous insects, assumes an interaction between two receptor types: green-sensitive, which contributes positively, and blue-sensitive, which contributes negatively. The preference displayed by many phytophagous insects for yellow over green stimuli can be explained by this model [50].
A similar mechanism based on receptor opposition was recently proposed for tsetse flies. In these hematophagous insects, the attraction to blue visual bait might be due to an interaction with a positive contribution by a blue-sensitive photoreceptor against photoreceptors sensitive to green-yellow-UV, which contribute negatively [18,36]. Moreover, based on the four sensitivity peaks of the tsetse fly, the number of specimens attracted by different colour panels correlated positively with the blue colour band and with reflectance at 460 nm (blue wavelengths), whereas the correlation was negative in the green colour range and for reflectance at 520 nm (green wavelengths) [36]. Glossinidae and Hippoboscidae share phylogenetic, morphological, and some behavioural features [51,52]; thus, as already supposed, deer keds are thought to display the same visual ecology as seen in the Glossina species. As a consequence, catches of L. fortisetosa by different coloured panels might have been determined by different reflectance in relation to the sensitivity of fly photoreceptors. The blue panel had a distinct reflectance peak at about 450 nm, which could be in the proximity of the sensitivity peak of L. fortisetosa. The green traps showed a reflectance maximum at about 500 and 550 nm, while the yellow ones displayed a "cut off" spectrum between 450 and 520 nm. The lower attractiveness of the yellow panels might be explained by an opposition mechanism of the photoreceptor sensitive to the yellow-green band as proposed for tsetse flies.
In a natural environment, green leaves have a green reflectance peak at 555 nm due to chlorophyll. They contrast with groups of objects that do not reflect the same wavelength (for instance, fruit and flowers). Moreover, grey-red surfaces (such as bark, soil, and animal) are also present, and these have a reflectance that increases gradually as a function of wavelength [53]. For glossinids, it has been suggested that the contrast of blue against the green-yellow reflectance of vegetation could represent a "non-vegetation" stimulus that induces flies to move towards a more feasible stimulus likely coming from a host. Moreover, the stronger attraction for blue could be explained by hypothesizing an important role of the shadow in creating contrasts: the shaded areas where flies rest, or darker areas on the bodies of potential hosts can appear as bluish patches that contrast with the background [18,36].
In different species of Glossinidae, a negative contribution of the green-red and ultraviolet wavelength in attracting tsetse flies was highlighted [18]. In L. fortisetosa, the number of individuals caught by green and red traps was higher than for yellow traps but was lower than for blue traps. In the red and green colour panels, however, the number of flies was similar to that found in black and transparent traps. Spectrophotometric analysis of the green panels showed a nearly constant reflectance at different spectrum bands with a maximum 20% at about 500 and 550 nm. On the other hand, the red panels showed a cut-off transition that varied from a minimum value of 535 to a maximum of 635 nm; the lower wavelength chromatic bands were therefore excluded (blue-green region, 410-520 nm).
Because deer have crepuscular and nocturnal activity, they move and find refuge inside the woodland during the day but feed and rest in open areas in the evening and during the night [54,55]. As proposed for the genus Glossina, we hypothesized that L. fortisetosa flies also show a colour response depending on the perception of the host against the background of green foliage or other background material. Alternatively, winged adults might prefer blue targets because they resemble the shaded areas where hosts rest or the shaded zones of the host body where insects can more easily carry out their parasitic action.

Trap Activity at Different Positions
As reported in Figure 6, the blue traps caught a higher percentage of insects than those of other colours for all positions. Regarding the three tested series positions, we found that hippoboscids were not caught in the same numbers in the different locations. In fact, in position 1, located at the border between the forest and the open area, captures were significantly higher than those in the other two, both of which were positioned among the trees. The more abundant captures of L. fortisetosa at the boundary between the woods and clearing were also supported by observations conducted in Japan [56]. These outcomes showed that the environment seemed to affect parasitic attraction, which was higher in position 1 probably due to the ecology of both fly and host. Since L. fortisetosa adults spend their entire lives in the fur of their host, reproducing and laying larvae, the pupae are more likely to fall from the host body to the ground in areas where the deer stay and repose. Because newly emerged winged adults are unlikely to be able to fly long distances, as reported for L. cervi [2], they are probably more abundant at emerging sites. Moreover, as already shown, the series were differently orientated with the first east-west oriented (trap sides with north-south exposure) and the second and the third north-south oriented (trap sides with east-west exposure). In Glossinidae, as already mentioned, ultraviolet radiation was negatively associated with fly catches [19,36]. Surface UV reflection varies with solar zenith angles and surface type and orientation [57,58]. The different orientation and exposure of the rows in the three positions may have determined a different reflectance that probably induced a lower presence of L. fortisetosa in areas with higher ultraviolet reflectivity. This marked difference in captures in the three positions with a total number of 1013 flies in the first compared with the 162 and 77 insects in positions 2 and 3, respectively, may have also been influenced by wind direction and speed. In the end, the flies showed a similar preference to colours in all three series, with blue being the most attractive and yellow the least.

Pattern of Lipoptena fortisetosa Captures in the Sampling Period
Concerning the number of L. fortisetosa adults caught during the sampling experiment (from 15 July to 1 October), it is important to highlight the remarkable peak in August, which might be due to adult emergence from pupae laid during the previous year. The slight increase in the number of flies caught in September might be attributed to a second generation from a small number of specimens originating from earlier emerged adults that quickly found a suitable host. Larvae that pupated in early summer could have completed their development and given rise to another emergence of adults favoured by the high temperature. In the Schignano area, the average temperature from mid-July to mid-September was 23.7 • C [59], while it was 14.1 • C from mid-September to the end of October. Lower temperatures of this latter period probably induced pupae to enter diapause until the following year [60]. Our experiment represents the first field-monitoring survey of L. fortisetosa and allows us to conclude that it seems to be multivoltine, as has been reported for other European countries [5,61] and for the area of origin [62].

Conclusions
Our research provides insight into the life cycle and basic visual ecology of L. fortisetosa, which may be exploited in the development of strategies for its monitoring and control and opening a new path to acquire fundamental knowledge on these ectoparasites.
Our experiment showed a peak presence of L. fortisetosa winged adults in mid-August, suggesting the possibility that, in Italy, this species could produce more than one generation per year.
More interestingly, we provide evidence that L. fortisetosa exhibits a preference scale for colours, with blue being the most attractive and yellow the least attractive. The colour ranking displayed by L. fortisetosa suggests that this species is able to discriminate colours and uses visual stimuli over short distance. These results could help in the design of traps to monitor and reduce the populations of this parasitic, hematophagous fly which, in some periods, can reach a very high density, causing annoyance by biting humans in natural areas. The way these flies locate a host remains a topic of active investigation, and further observations are needed to better define the complex stimuli that govern this behaviour. In particular, the role played by odours at medium and long distance should be clarified. For instance, we can state that L. fortisetosa, contrary to what is reported for L. cervi, does not passively search for a host by waiting for an animal to pass by; rather, it actively flies in search of one. This is supported by our capture of winged adults by a sweeping net only when they were flying, and not when they were resting on vegetation.
The noticeable consistency in some aspects of host attraction among different groups of hematophagous insects is surprisingly high and may suggest a general similar set of needs and tendencies in blood-feeding ectoparasites. Particularly, the convergence in visual stimuli and colour preference ranking is evident: dark colours, especially blue, red, and black, are often attractive for parasites. The similar colour preference for different groups of ectoparasites made us wonder why some hematophagous dipterans are particularly attracted by blue and seem to be relatively unattracted by yellow. Future observations on different blue wavelengths and reflectance are needed to set appropriate and effective traps for this parasite.