Study Species. E. tricuspidata is one of the most common flower-visiting spiders and is widely distributed. We conducted our study by E. tricuspidata collected from Shahu Park in Wuhan City (30° 35′ N, 114° 20′ E), which is close to Hubei University. After collection, we transferred the spiders into individual glass tubes (diameter 2 cm, length 6 cm). The spiders were maintained in a light incubator laboratory under controlled environmental conditions (temperature: 25 ± 1℃; relative humidity: 50%‒70%; light regime: 14 h: 10 h, with lights coming on at 08:00). Every spider was fed with 10 fruit flies (Drosophila melanogaster) or 2 houseflies (Musca domestica) in turn every 3 days and watered daily. All of the spiders used in the experiments were adult females.
Volatile Scent Compounds. To determine whether and which volatile scent compounds attract E. tricuspidata to tulip flowers, we selected eight main floral volatile scent compounds of tulips based on literature (Knudsen et al. 2006; Oyama-Okubo and Tsuji 2013): four monoterpenes (limonene (5989-54-8), β-ocimene (13877-91-3), linalool (78-70-6), and α-pinene (80-56-8)) and four benzenoids (benzaldehyde (100-52-7), benzyl-alcohol (100-51-6), methyl-salicylate (119-36-8), and 2-phenylethanol (60-12-8)), and conducted dual choice behavioral assays for the eight compounds separately. All compounds were purchased from Shanghai Aladdin Biochemical Technology Co., Ltd. The CAS numbers of compounds are presented in the brackets.
Volatile Scent Compounds Preference Experiment. Dual choice behavioral assays were conducted using a Y-tube olfactometer (Fig. 1). In the olfactometer device, there is a Y-tube (length 10 cm, diameter 1 cm, angle 45°) as the odor arm, control arm, and test arm, two glass tubes (length 10 cm, diameter 2 cm) as the odor chamber and control chamber, two flowmeters (Lzb-3wb, Changzhou Shuanghuan Thermal Instrument Co., Ltd, China) for controlling the airflow rate, and an air pump (Sensen Air Pump, CT-202, China).
Each of the eight scent compounds was tested against an odorless control by the Y-tube olfactometer (n = 30 for each compound). Scent compounds were emitted from filter paper or cotton balls, which were soaked in 7 µL/mL synthetic scent compound solution and subsequently dried for 10 min to reach constant compound emission rates (Knauer et al. 2018), meanwhile, the control filter paper or cotton balls were also soaked in solvent and dried for 10 min. For water-insoluble scent compounds, anhydrous ethanol was used as the solvent.
The dual choice behavioral assays were carried out at 22 ± 2 ℃. The sample and the control filter paper or cotton balls were put in the odor and control chambers, respectively. A spider was randomly selected and put into the end of the test arm then we turned on the air pump at 150 mL/min for 10 min. After that, we started the trial by removing the breathable metal screen for the spider to make the choice to move to the odor arm or the control arm freely. The time the spider spent getting into the odor or control arm (hereafter ‘choosing time’) was recorded. If the spider stayed in the odor or control arm for at least 60 second after arriving, we regarded the spider made a choice. We ended the trial if the spider did not get into either arm within 5 min; pervious experiments showed that spiders would not move to any arm if they did not make a choice within 5 min. During the experiment, spiders should not be disturbed, otherwise they would show nervous immobility and escape behavior. We replaced it with another individual randomly if the spider was disturbed.
Every spider was used once at most. After each choice by a spider, the Y-tube olfactometer was disassembled and cleaned with 80% ethanol and water, and the positions were exchanged in the next trail.
Statistical Analyses. SPSS 20 was used for all data analyses. We analyzed the preference choice results using chi-square goodness of fit test. The expected value was 1:1. As for the choosing time, we used the Shapiro-Wilk test to analyze the normality for each compound and control. Student's t-test was used to analyze the data with normal distribution, and the Mann-Whitney U test was used to analyze the data that were not normally distributed. Statistical analyses charts were made using the software GraphPad Prism 7.