Attraction effect of different colored cards on thrips Frankliniella intonsa in cowpea greenhouses in China

The flower thrips Frankliniella intonsa (Trybom) is one of the most economically important pests in cowpea greenhouses in China. Widespread pesticide resistance of thrips and the negative environmental effects limit the application of pesticides for thrips control. Two commercial cowpea greenhouse experiments were designed to determine the color preference of F. intonsa to colored cards, including white, pink, pale green, light yellow, powder blue, crimson, yellow green, deep sky blue, dark slate blue, dark orange, medium orchid, gold, and black. Clear pieces of plastic were used as the control cards. Additionally, the effects of placement height and orientation (cardinal direction) of the cards were also studied. Both greenhouse trials showed that white cards were significantly more attractive to F. intonsa than the other 13 card colors, followed by deep sky blue cards. White or deep sky blue cards placed low to the ground were found to be most attractive to F. intonsa. Orientation of the colored cards also affected the attractiveness to F. intonsa. The results indicate that white sticky cards were significantly more attractive to F. intonsa than blue cards and therefore can be recommended to monitor F. intonsa population densities and to control them in cowpea greenhouses as part of integrated pest management programs.


Results
Effect of card color on thrips. The numbers of thrips attracted to the cards varied among the 13 colors (Figs 1 and 2). The control (clear) cards were unattractive to thrips (less than 2 and 1 thrips per card, respectively) in both greenhouse trials (Tables S1 and S2). Among the 13 colored cards, the white cards had the strongest attractive ability, attracting up to 61 thrips per card in trial I and 474 thrips per card in trial II (Tables S1 and S2). White cards were significantly more attractive than the other color in both trials at 2 h, 4 h and 6 h after the start of the experiment (P = 0.05) (Figs 1 and 2).
In greenhouse trial I, the white cards had the strongest attractive ability to thrips, followed by deep sky blue, powder blue, medium orchid, gold and the other color after 2 h (Fig. 1). Furthermore, the number of thrips attracted to the deep sky blue cards was similar to the powder blue cards but was significantly higher than the other colored cards (P = 0.05) (Fig. 1). After 4 h, the numbers of thrips attracted to the white cards increased, while the number of thrips attracted on the deep sky blue or powder blue cards both decreased (Fig. 1). Furthermore, the number of thrips attracted to the deep sky blue cards was similar to the powder blue cards, but significantly higher than the other colors (P = 0.05) (Fig. 1), which all showed no significant difference from the control (clear) cards in attracting thrips (P = 0.05) (Fig. 1). After 6 h, white cards still showed the strongest attractive ability, followed by deep sky blue (Fig. 1), and deep sky blue cards attracted significantly more thrips than each of the other colors, except for white (P = 0.05) (Fig. 1).
In greenhouse trial II, white cards also had the strongest attractive ability (303 thrips per card) after 2 h, followed by deep sky blue, powder blue, medium orchid and the other colors (Fig. 2). Furthermore, the number of thrips attracted to deep sky blue cards (161 thrips per card) was significantly higher than to powder blue cards (78 thrips per card) (P = 0.05). In addition, the number of thrips attracted to the powder blue cards was similar to medium orchid cards, but significantly higher than the rest of the colors (P = 0.05) (Fig. 2). After 6 h, the numbers of thrips attracted to the white, deep sky blue, powder blue, and medium orchid cards all increased, except for the powder blue cards (Fig. 2). After 4 h, the number of thrips attracted to deep sky blue cards was similar to powder blue cards, but significantly higher the other colors (P = 0.05) (Fig. 2). Except for white, deep sky blue, powder blue and medium orchid, there was no significant difference between each of the other colored cards and the control (clear) cards (P = 0.05) (Fig. 2). After 6 h, the deep sky blue cards showed significantly stronger attractive ability than the powder blue cards (Fig. 2). In addition, the number of thrips attracted to the powder blue cards was similar the medium orchid cards but was significantly higher than the other colors (P = 0.05) (Fig. 2).
Effect of card placement height on thrips. The white and deep sky blue cards, as the two most attractive colors, were selected to evaluate the effect of card placement height on thrips. The number of thrips attracted to the white or deep sky blue cards differed with card placement height (Table 1). For the white cards, the number of thrips attracted at the low position was always significantly higher than that at the middle or high position  (P = 0.05), except for the middle position in trial I at 6 h. For the deep sky blue cards, the number of thrips attracted at the low position was also always significantly higher than at the middle or high position (P = 0.05), except for the middle position in trial I at 4 h. In both trials, whether the cards were white or deep sky blue, there was no significant difference between the middle position and the high position at 2 h, 4 h and 6 h (P = 0.05) ( Table 1).

Effect of card cardinal direction on thrips.
The white and deep sky blue cards, as the two most attractive colors, were also selected to evaluate the effect of card orientation (cardinal direction) on thrips. The number of thrips attracted to the white or deep sky blue cards varied with card orientation ( Table 2). For white cards in trial I at 2 h, there were more thrips attracted to the side facing to the south or east than to the side facing north or west (P = 0.05). However, the significant difference disappeared at 4 h or 6 h. There was no significant difference between cardinal directions in trial II (P = 0.05), except for the side facing east at 4 h ( Table 2). For the deep sky blue cards in trial I at 2 and 4 h, there were more thrips attracted on the side facing to the east than to the side facing west (P = 0.05). However, there were more thrips attracted to the side facing south than to other cardinal directions in trial II (P = 0.05).
As one card always has two sides, the number of thrips on the two sides was summed for further analysis, for example, east-west and south-north (Table S3). For white cards, there was no significant difference in the number of thrips attracted to east-west versus south-north in both greenhouse trials, except for trial II at 2 h. For the deep sky blue cards, there was also no significant difference in trial I, but the number of thrips attracted to south-north facing cards was significantly higher than east-west facing cards in trial II at 2 h, 4 h and 6 h (P = 0.05) (Table S3).

Trials
Color Height

Discussion
Blue was reported to be the most attractive color to F. intonsa in a balsam pear field 15 and a strawberry field 16 , followed in attractiveness by white, but there was no significant difference between blue and white in the balsam pear field 15 . However, white cards were found here to have a significantly stronger attractive ability than blue cards (deep sky blue or powder blue) for F. intonsa in our two cowpea greenhouse trials. Furthermore, the attractiveness to F. intonsa varied among different blue colors in the two trials, for example, deep sky blue, powder blue and dark slate blue (Figs 1 and 2). It may be possible to find another blue color except for the three blue colors tested here (deep sky blue, powder blue and dark slate blue) that was as attractive as white to F. intonsa. However, the greenhouse trials here indicated that white is the best color for attracting F. intonsa, rather than blue or yellow cards in cowpea greenhouses. In our two cowpea greenhouse trials, the cards at the low position were usually found to attract more thrips than at the middle or high positions (P = 0.05) ( Table 2), which could be attributed to thrips behavior of entering the soil for pupation 17 . Furthermore, both of the two greenhouse trials were carried out at noon, so thrips perhaps gathered at lower positions of the cowpea plants to escape the strong sunlight due to the concealment behavior 18 . This may be another reason for attracting more thrips at the low position.
In our two cowpea greenhouse trials, the main species of thrips was identified to be F. intonsa in each case. Indeed, there are many thrips species infesting different crops in the field, but different species of thrips are attracted to different colors 14,19 . To effectively monitor thrips population densities and control different thrips species, first testing different colored cards to screen for optimal color is necessary before applying colored cards in the field. To enhance the pest control of F. intonsa, pheromone could be used in combination with white sticky cards 20 .
Additionally, in order to successfully control F. intonsa with IPM, other biological control agents should also be considered 21 , for example, silver stripe mulching film 22 , predatory mites 23 , and entomopathogenic fungi 18,24 . No beneficial insects were present during our two greenhouse trials, so the potential effect of colored cards on beneficial insects (e.g., lacewings) should also be considered in future work 25 .
In summary, white cards showed the strongest attraction ability for F. intonsa among the 13 different colored cards in each of two cowpea greenhouse trials. However, more detailed work on potential combinations with pheromones or other biological agents (e.g., predatory mites, entomopathogenic fungi or others) are necessary before white sticky cards can be recommended as an efficient method to control F. intonsa in cowpea greenhouses in China.
Colored card selection. Thirteen colors were selected from the Encycolorpedia (http://encycolorpedia. com/) for making colored cards (white, pink, pale green, light yellow, powder blue, crimson, yellow green, deep sky blue, dark slate blue, dark orange, medium orchid, gold, and black). Cards of all of the above colors were printed on A4-size paper (297 mm × 210 mm) and sealed with two pieces of clear laminating film. Clear pieces of plastic with the same size were used as the untreated control. The RGB values of the above cards are listed in Table 3.
Experimental design. The thirteen colored cards and clear controls were employed as randomized blocks with four replicates each in trials I and II in July, 2017. Two cardinal direction (east-west, and south-north) and three placement heights (0 cm, 80 cm, and 160 cm) were studied at the same time. Every plot area was designed in  15 m 2 with 6 same colored cards. Six cards of each color were employed at each of six positions (east-west/0 cm, east-west/80 cm, east-west/160 cm, south-north/0 cm, south-north/80 cm, south-north/160 cm) and were placed randomly in each plot. The number of thrips attracted by both sides of the cards were counted at 2 h, 4 h and 6 h after the placement of cards. The whole experimental design was briefly descripted in Fig. 3, and all of the photos in Fig. 3 were taken in the greenhouse trials and the laboratory by the authors.
Statistical analysis. The average number of thrips per card, including both sides, for each color in each plot were used to evaluate the attraction effect of the different colors. The average number of thrips per card for each height by two different cardinal directions (east-west, and south-north) in each plot were used to evaluate the attraction effect of card height. The average number of thrips attracted per side of each card for each of the four cardinal directions and by three different placement heights (0 cm, 80 cm, and 160 cm) in each plot were used to evaluate the attraction effect of card orientation. All of the treatments were performed with four plot replicates at each of the two trials. Data for thrips populations were transformed as necessary for statistical analyses (square root transformations for small numbers [<100] and log10 for large numbers [>100]), but all data are reported as untransformed values. Data were analyzed using ANOVA with SPSS (version 22.0 for Windows). Significant differences among means were tested with Fisher's LSD at P = 0.05 26,27 .