3.3 Seed symbiotic and asymbiotic cultures
More germinated seeds with low developmental stages (stage 2 and stage 3) were found after self-pollination (Figure 3A) and nectar addition (Figure 3B) compared with that after cross-pollination (Figure 3C) and natural conditions (Figure 3D) under the conditions of 30-day seed symbiotic culture. There were more seedlings with high developmental stages (stage 5 and stage 6) after cross-pollination (Figure 4C) and natural conditions (Figure 4D) compared with that after self-pollination (Figure 4A) and nectar addition (Figure 4B) under the conditions of 70-day seed asymbiotic culture.
There was no significant difference in seed germination among the four treatments under the conditions of 60-day symbiotic and 70-day asymbiotic cultures (all p > 0.05) (Figure 5A). Average fresh weight of a single germinated seed after self-pollination and nectar addition was significantly lower than that after cross-pollination and natural conditions (all p < 0.01). Moreover, there were significant differences in average fresh weight of a single germinated seed between self-pollination and nectar addition (p <0.05) but not between cross-pollination and natural conditions (p>0.05) (Figure 5B). The results of average developmental stages were similar to that of average fresh weight of a single germinated seed (Figure 5C).
Under the condition of symbiotic culture, percentage of germinated seeds with high developmental stages (stage 5 and stage 6) after cross-pollination and natural conditions was significantly higher than that after self-pollination and nectar addition, while that with low developmental stages (stage 2 and stage 3) was significantly lower (all p < 0.01) (Figure 5D; Table 2). Differences in percentage of germinated seeds with developmental stage 2-6 did not reach significant level between cross-pollination and natural conditions (all p > 0.05) (Figure 5D; Table 2). There were significant differences in percentage of germinated seeds with developmental stage 3 and stage 6 between self-pollination and nectar addition (all p < 0.01) (Figure 5D; Table 2).
Under the condition of asymbiotic culture, percentage of germinated seeds with high developmental stages (stage 6) after cross-pollination and natural conditions was significantly higher than that after self-pollination and nectar addition, while that with low developmental stages (stage 3 and stage 4) was significantly lower (all p < 0.01) (Figure 5E; Table 2). In most cases (except stage 4), percentage of germinated seeds at different developmental stages showed no significant difference between cross-pollination and natural conditions (all p > 0.05) (Figure 5D; Table 2). There were significant differences in percentage of germinated seeds with developmental stage 2, stage 4 and stage 6 between self-pollination and nectar addition (all p < 0.01) (Figure 5D; Table 2).
Discussion
Like the other orchids (Johnson & Edwards, 2000; Johnson et al., 2004), pollinarium bending obviously exists in Papilionanthe teres , and was reached in 62.43 ± 22.97 s (N = 30). Total time per visit under natural conditions and after nectar addition were significantly lower and higher than that of pollinarium bending, respectively (Figure 1A), suggested that the two conditions could delay and accelerate touching between pollinia and stigma, respectively. Nectar addition significantly increased the number of flowers visited in a single inflorescence and pollinator staying time in a single flower (Figure 1B). Besides, the proportions of full seeds and seed viability after nectar addition were significantly lower than that under natural conditions (Table 1; Figure 5). These results indicated that P. teres is similar to most deceptive orchids that avoid geitonogamy by the two mechanisms of deception and pollinarium bending (Darwin, 1877; Dafni & Ivri, 1979; Johnson & Edwards, 2000; Johnson et al., 2004), and deception contributes to pollinarium bending functioning in the aspect of anti-selfing.
Both self-compatibility and inbreeding depression can be checked by pollination experiments. There was no significant difference in fruit set between self and cross-pollination in the earlier stage of fruit development, while self-pollination led to serious fruit abscission in the late stages (data not shown). Compared with cross-pollination, there were more shriveled seeds, smaller weight of a single fruit and fewer seeds per fruit after self-pollination (Figure 2A, C; Table 1). These results indicated that self and cross-pollination result in different female reproductive success. Moreover, self-pollination seriously affects the development of fruits and seeds of P. teres .
Most orchids are self-compatible species, and the seed set and seed quality are usually greatly reduced after self-pollination (Tremblay et al., 2005). Compared with cross-pollination and natural conditions, self-pollination and nectar addition had more seeds with low developmental stages under the conditions of 30-day (Figure 3A, B), 60-day (Figure 5C, 5D; Table 2) symbiotic culture and 70-day asymbiotic culture (Figure 4A, B; Figure 5C, E; Table 2), suggesting that seed viability was greatly reduced after self-pollination and nectar addition. There were significant differences in average fresh weight of germinated seeds (Figure 5B), average developmental stage (Figure 5C), proportions of seeds with high developmental stage (stage 6) (Figure 5D, E) and shriveled seeds (Table 1) between self-pollination and nectar addition under the conditions of symbiotic and asymbiotic cultures, suggesting that nectar addition partially results in geitonogamy. It is important for orchids to reach higher developmental stage faster under natural conditions to enhance its survival and resistance to disadvantageous environmental and climate conditions (Smith et al., 2007; Stewart, 2008). Based on this conclusion, and combined with our study results, it is confirmed that germinated seeds after cross-pollination and natural conditions have higher ability to adapt to the environment than that after self-pollination and nectar addition.
Seed morphology (Figure 2C, D), and parameter values of fruits and seeds (Table 1) showed no significant differences between natural conditions and cross-pollination. Besides, there was no significant difference in average fresh weight of germinated seeds (Fig. 5B), average developmental stage (Figure 5C), proportions of seeds with different developmental stages (Figure 5D, E) between natural conditions and cross-pollination under the conditions of symbiotic and asymbiotic cultures. Taken together, these results suggested that P. tereshas high outcrossing rate under natural conditions. Some case studies confirmed a negative correlation between inbreeding depression and selfing rate (Husband & Schemske, 1996), indicating that the lower the selfing rate, the more obvious the inbreeding depression. Thus, if deceptive orchid species regularly experience high levels of outcrossing, a higher cost of selfing is expected in deceptive species than in rewarding ones (Jersáková et al., 2006). Compared with cross-pollination and natural conditions, there were more shriveled seeds (Figure 2), lighter weight of a single fruit and seeds per fruit (Table 1), and lower seed viability (Figure 5; Table 2) after self-pollination and nectar addition under the conditions of symbiotic and asymbiotic cultures, indicating that both self-fertilization and geitonogamy result in inbreeding depression. These results further confirmed that P. teres has high outcrossing rate under the natural conditions.
Nectar addition experiments along with field observation were used in this study to comparatively analyze the differences of pollinator visit behavior, and the results confirmed thatP. teres decreased geitonogamy by the two mechanisms of deception and pollinarium bending. Deception contributes to pollinarium bending playing the function of anti-selfing in P. teres . Besides, the fruits and seeds from four sources (self-pollination, cross-pollination, nectar addition and natural conditions) were used to compare the differences in the characteristic parameters of seeds and fruits, and check the seed viability by symbiotic and asymbiotic culture techniques, which indirectly proved that P. teres had high level of genetic load. Both self-fertilization and geitonogamy resulted in inbreeding depression, such as high proportion of shriveled seeds and lower developmental speed of germinated seeds. These conclusions support the outcrossing hypothesis that ecological significance of P. teresdeception is to promote outcrossing and improve the ability of the offspring to adapt to the environment.