Chapter 4 Discussion
4.3 Reproductive features in D. arisanensis
Several ecological features found in D. arisanensis, including perennials, unspecialized pollinators, flesh fruits and island habitats, had been suggested to be correlated with dioecism (Bawa, 1980; Renner, 2014). Strictly dioecious species are thought to be obligate outbreeding, since one individual cannot reproduce sexually without its complementary gender. Due to this non-independence, the correlated ecological factors had been discussed principally from the aspect of promoting reproductive opportunities, together with gender optimization and effective colonization.
The dioecy in D. arisanensis is not obligate but actually leaky, since either a female or a male could set fruits on its own through apomixis or partial feminization respectively.
Nonetheless, these factors might still be influential in the predominantly outbreeding system, and the probable interactions are briefly discussed below.
Perennial life-form has a strong correlation with the occurrence of dioecy, and actually only several cases of dioecious annuals had been reported (Bram & Quinn, 2000;
Decker & Pilson, 2000). For a newly-colonizing individual, it may take years to meet with the coincident arriving of opposite sex, thus the perennial-form would be more stable in dioecy than annual. Moreover, perennial trees with simultaneous large floral display would suffer greater pressure of self-pollination between neighboring flowers (i.e.
geitonogamy), which could be a selective force driving gender divergence (discussed further in section 4.4).
Adaptations to various unspecialized pollinators may also benefit in colonizing new
habitats, where pollination service could be more unreliable. Besides, it has been suggested that specialized pollinators might have a better ability to discriminate between the differential rewards of males and females in a dioecious species, which would lead to an unequal flower-visiting rate and have negative effect on total reproductive success (Charlesworth, 1993).
Daphne fruits are usually fleshy and berry-like (Wang et al., 2007), which is a character usually associated with zoochory in long distance dispersal. The maternal burden of setting such fruits had been suggested to associate with the evolution of differential resource allocation and the further gender divergence (Bawa, 1980).
The occurrence of dioecy in islands seemed to be a controversy to Baker’s law (Baker, 1955), which stated the relative difficulty of colonization after long-distance dispersal with a non-self-fruiting individual (self-incompatible or unisexual). A complex integration among several factors may account for the necessary advantages, including the mentioned longevity, long-distance and multi-unit dispersal via zoochory, utilization of unspecialized pollinators, together with the probable strong selection for out-breeding in islands and the leakiness in dioecy (Baker & Cox, 1984; Emerson et al., 1996).
Leaky dioecy is characterized by the occasional presence of opposite sex or bisexual flowers within a unisexual individual (Baker & Cox, 1984). The leakiness could be either a remnant during gender divergence or a consequence of further selection with adaptive significance. As the majority of dioecious species are self-compatible (Charlesworth, 1985; Thomson & Barrett, 1981), leakiness could bring about small-scale reproductive assurance, and it would be especially beneficial during long-distance colonization. In D.
arisanensis, male plants could produce several feminized flowers, which were capable of self-pollinating and siring viable seeds. Though no fertile male organ was found in the female plants, they could still reproduce via apomixis and contribute to a similar
consequence of leakiness. Since D. arisanensis is endemic and widely distributed in Taiwan, it would be challenging to find suitable isolated populations for testing the dispersal model. Exploited areas like Meifeng might have given the opportunity for secondary colonization, and the population genetics would be informative on the past colonizing-pattern (see section 4.2). It is noteworthy that the occurrence of apomixis in dioecious species is relatively uncommon (Dupont, 2002; Richards, 1990). In this study, the hermaphroditic D. kiusiana var. atrocaulis also displayed apomictic reproduction in the emasculation-unpollination treatment (Table 3-5), but further examination in the ontogeny of embryos is needed to clarify the underlying mechanism and to exclude the possibility of contamination problem during manipulation. Rare fruit-set from unpollinated flowers were also reported in the congeneric gynodioecious D. laureola (Alonso & Herrera, 2001), both in bisexual flowers and female flowers, but the authors simply regarded the results as contaminations. If apomixis was common in Daphne, it could be a preadaptation to the gender divergence in D. arisanensis. However, given that the flowers of Daphne are easily self-pollinated (further discussed in section 4.4), the advantage of apomixis seem to be redundant in other hermaphroditic species. A broader-scale analysis is needed to clarify the existence of apomixis and the correlation with gender divergence in the whole genus.
The sterile but morphologically complete gynoecia in male flowers made them a bisexual appearance, thus the sexual system in D. arisanensis could be termed as cryptic dioecy (Mayer & Charlesworth, 1991). In cryptic dioecy, the rudimentary organ may represent an evolutionary remnant without adaptive significance, or have its specific function in pollination. In the former situation, a long time for selective pressure may be needed to promote degeneration of the sterile sex-organs, and the complicated genetic control of floral development may bring difficulties in reducing the non-functional floral
part without affecting others (Mayer & Charlesworth, 1991). The later functional explanation was mainly met with the presence of male organs in female flowers. Pollens and conspicuous anthers may be still important to maintain the attractiveness to pollinators in females, thus the retention of sterile androecia have a positive effect on pollination. While being adapted to nocturnal-moth-pollination, there was no need to produce sterile pollens as rewards in the female flowers of D. arisanensis. Generally, no obvious contribution to promoting pollen-dispersal could be found on the rudimentary gynoecia of male flowers, Mayer and Charlesworth (1991) suggested that reducing visual differences between the males and females would be helpful in maintaining equivalent visiting rate. In the case of D. arisanensis, many pollens would even be intercepted by the self-stigma, resulting a discount of potentially dispersible pollens (further discussed in section 4.4). Another possible functional perspective is that the retention of gynoecia could be essential in expressing leakiness (i.e. feminization in D. arisanensis), but the potential benefit seem to be inequivalent to the cost.
The fleshy fruits and the seeds of Daphne were supposed to be dispersed mainly via endozoochory of birds, which had been recorded in D. gnidium (Herrera, 1981) and D.
laureola (Brickell & Mathew, 1976). The developmental and functional role of the aril-like membrane surrounding the seed coat was not clear yet. Due to its fragility, it probably would be destroyed when passing through the digestive tract of a disperser. It had been reported that pulp may inhibit seed germination in other endozoochorous species, and removing the pulp could be an indicator of successful dispersal and a suitable environment (Traveset et al., 2007). I was unable to test this hypothesis in D. arisanensis because the seeds easily got moldy with retaining the moist membranes for months. The seeds exhibited various and continuous lengths of dormancy. Low-temperature stratification seemed to have no synchronizing effect. Mechanical damage on the
crustaceous seed coat significantly boosted the germination curve, but some of the seeds still remained dormant for two-three months. Thus, both physical and physiological effect underlay the dormancy pattern. If dispersal by endozoochory does not give a stronger effect than manipulated-scarification, the naturally dispersed seeds would exhibit various lengths of dormancy in soil. Also considering the variable flowering seasons among populations, the emergence of seedlings may be irregular all year round. In summary, the germination strategy of D. arisanensis is likely a seed bank formation (Roberts, 1981).