2.2 Manual control experiments
In order to obtain P. teres seeds from different sources, four types of manual control experiments including nectar addition, natural conditions, self and cross-pollination were conducted at the six study sites during the anthesis of P. teres in two consecutive years (2019-2020). All flowers at FG were selected to conduct nectar addition experiments. Due to partially overlapped antheses and the same pollinator of rewarding Thunbergia grandiflora and food-deceptiveP. teres , we mimicked the open flowers of T. grandiflorato produce artificial nectar. Spur of each newly-open flower of P. teres was supplied with artificial nectar (20 μl volume and 50% sugar content) and the surplus nectar was removed. Fresh nectar was re-added using a microsyringe every morning at 08:00 am until the flower wilted. Nectar addition did not result in floral damage and had no effect on the floral longevity. So there was no necessity to set control of water addition. In order to obtain natural fruit and seed set, all flowers and inflorescences at AR, TSRA and VG were maintained in natural states. Two flowers of each inflorescence at EPG and EMG were used to conduct self (The pollinia were from the same or another flower in the same inflorescence) and cross (The pollinia were from the other flower at EPG) pollination, respectively. All flowers used in hand pollination were bagged before the flowers opened and they were bagged again after pollination. In next April, mature fruits of the four treatments were collected for subsequent experiments.