|Home | About | Journals | Submit | Contact Us | Français|
The thermoregulatory flower of the Asian sacred lotus (Nelumbo nucifera) can maintain a relatively stable temperature despite great variations in ambient temperature during anthesis. The thermoregulation has been hypothesized to offer a direct energy reward for pollinators in lotus flowers. This study aims to examine whether the stable temperature maintained in the floral chamber influences the fertilization process and seed development.
An artificial refrigeration instrument was employed to cool flowers during the fertilization process and post-fertilization period in an experimental population. The effect of temperature on post-pollination events was also examined by removing petals in two field populations.
Treatments with low floral temperature did not reduce stigma receptivity or pollen viability in undehisced anthers. Low temperature during the fertilization period significantly decreased seed set per flower but low temperature during the phase of seed development had no effect, suggesting that temperature regulation by lotus flowers facilitated fertilization success. Hand-pollination treatments in two field populations indicated that seed set of flowers with petals removed was lower than that of intact flowers in north China, where ambient temperatures are low, but not in south China, confirming that reducing the temperature of carpels did influence post-pollination events.
The experiments suggest that floral thermoregulation in lotus could enhance female reproductive success by facilitating fertilization.
Favourable temperature is crucial for various phases of sexual reproduction in flowering plants, including pollen development, pollen transfer, stigmatic receptivity, pollen germination, pollen-tube growth, double-fertilization and seed development (Pigott and Huntley, 1981; Kjellberg et al., 1982; Young, 1984; Corbet, 1990; Stephenson et al., 1992; Kudo, 1995; Delph et al., 1997; Hedhly et al., 2003). To maximize reproductive success, flowering plants have evolved various strategies to maintain an optimal microclimate within flowers. In some plants, flowers possess dense pubescence and overlapping bracts to minimize heat loss and protect their flowers from cold (Meinzer and Goldstein, 1985; Miller, 1986; Tsukaya et al., 2002), whereas others elevate floral temperature by positioning flowers in a convenient position on the plant (Lu et al., 1992) or sun-tracking to absorb solar energy (Hocking and Sharplin, 1965; Kevan, 1975; Corbet, 1990). In contrast to those plants passively receiving energy from the sun during anthesis, thermoregulatory plants can actively produce a significant amount of heat to warm themselves and can precisely adjust the rate of heat production in relation to ambient temperature (Seymour and Schultze-Motel, 1997).
The role of thermoregulation in Asian sacred lotus, Nelumbo nucifera, as well as other homoeothermic species, such as Philodendron selloum (Nagy et al., 1972), Symplocarpus foetidus (Knutson, 1974) and Dracunculus vulgaris (Seymour and Schultze-Motel, 1999), has been interpreted as rewarding endothermic beetles (Schneider and Buchanan, 1980; Seymour and Schultze-Motel, 1996) or warming beetles to flight temperature and thus allowing them to depart carrying pollen (Burquez et al., 1987), and accelerating scent emission. Given that these species share an obvious cantharophily syndrome including a large chamber enclosed by petals, a large number of stamens and odour release, homoeothermic flowers have been considered as a warm shelter for pollinator activity (Seymour and Schultze-Motel, 1997). Heat production attracts pollinators, or facilitates their departure (Burquez et al., 1987), helping to remove pollen and so facilitating male reproductive success. Meanwhile, the stable temperature maintained in flowers can warm the sexual organs, which may affect the development of pollen and ovules, and facilitate fertilization and seed development. For example, solar tracking enhances sexual reproduction by increasing pollination, fertilization and seed development in Adonis ramosa (Ranunculaceae, Kudo, 1995) and by promoting pollen germination in the snow buttercup Ranunculus adoneus (Galen and Stanton, 2003). Although it is known that favourable temperature is crucial for post-pollination events, it remains unclear whether stable temperature is necessary for fertilization and seed development in thermoregulating plants.
Lotus flowers can physiologically regulate the floral temperature during anthesis by heat production via a large obconical receptacle, maintaining a relatively favourable temperature between 30 and 36 °C when the ambient temperature fluctuates between 10 and 45 °C (Seymour and Schultze-Motel, 1996, 1998; Seymour et al., 1998). The homoeothermy coincides with the period of stigma receptivity in lotus flowers (Seymour and Blaylock, 2000), suggesting that events at, or soon after, pollination can be influenced by floral temperature. Here, we investigate the effect of temperature on post-pollination events in hand-pollinated flowers by reducing flower temperature using an artificial refrigerator in an experimental population. It was found that stigma receptivity and pollen viability in undehisced anthers were not influenced by low temperature, whereas low floral temperature during the fertilization period significantly decreased seed production. To test whether similar effects occurred in the field, flower temperature was reduced by removing petals and a comparison was made of seed production after hand-pollination in two populations experiencing different ambient temperatures. The field hand-pollination treatments confirmed that the stable temperature was important in post-pollination events in lotus flowers.
Nelumbo nucifera Gaertn (Nelumbonaceae) is an emergent aquatic plant with a wide distributional range in the Old World throughout temperate and tropical Asia (Borsch and Barthlott, 1994). The bowl-shaped flowers are solitary and large, up to 25 cm in diameter, emitting intense odours during anthesis. Each flower has about 20–30 carpels in an obconical receptacle and 200–400 stamens (Ni, 1987; Hayes et al., 2000). Each carpel has a single ovule. Placentation is laminar, and the ovule is anatropous, bitegmic, crassinucellate and pendulous (Gupta and Ahluwalia, 1977). Lotus flowers are protogynous, and anthesis of one flower lasts 3–4 d. The protruding stigmas of flowers on the first day of anthesis (hereafter termed D-1) are receptive and covered with a mucilaginous secretion, whereas undehisced stamens are tightly pressed by erect petals around the receptacle and are not accessible (Schneider and Buchanan, 1980). The male stage starts from the second day (D-2) of anthesis, as numerous stamens are revealed, bearing abundant pollen. Notable thermoregulation occurs in D-1 flowers and is maintained throughout the night when the flower petals close (Seymour and Schultze-Motel, 1996, 1998).
To test the effect of low floral temperature on post-pollination processes, a refrigerating instrument was constructed. In an experimental population at Wuhan Institute of Vegetable Science, Hubei Province (29°58′ N, 113°41′ E) during 2007 and 2008, a heat-insulated foam box (10 × 10 × 10 cm) with a circular hole (5 cm in diameter) at the bottom in the centre, allowing a flower with petals removed to be inserted, was used to simulate a real floral chamber. To lower the internal temperature, some bags of ice were put inside the box without touching the receptacle and stamens of the petal-removed flower. Because heat production by lotus flowers does not depend on the light cycle (Seymour et al., 1998), a cover was added on the box in order to improve the reduction of internal temperature during each refrigeration treatment. A Delta TRAK FlashLink data logger (http://www.deltatrak.com) was used to monitor temperature fluctuations in the interior of the refrigerated and control flowers.
In order to examine the effect of low temperature on stigma receptivity, two treatments were applied to D-1 flowers commencing in the morning: (1) refrigeration for about 11 h from 0600 h to 1700 h, and then pollinated with normal pollen (see below for different pollen types); (2) no refrigeration during the daytime, and then pollinated at about 1700 h with normal pollen. The period from pollination to fertilization (hereafter termed the fertilization process) is reported to last 6–8 h in lotus flowers (see Yan, 1986). To explore the effects of temperature fluctuation on the fertilization process, and/or post-fertilization development, D-1 flowers were hand-pollinated in the morning and then subjected to four additional treatments: (3) refrigeration immediately from 0600 h to 1700 h (daytime refrigeration); (4) no refrigeratation until 1700 h, but then refrigerated throughout the night (nocturnal refrigeration); (5) petals were removed around 0600 h without refrigeration; and (6) no refrigeratation or removal of petals (control). In order to test the effect of low temperature on pollen viability in undehisced anthers, we pollinated normal D-1 flowers with pollen shed from flowers previously given treatments (3) and (4). Melting ice bags were replaced around noon with newly frozen ones to ensure that the temperature in the box was maintained below 30 °C.
The effect of floral temperature on reproductive success was further examined by removing petals in two field populations of lotus, at Lantian and Mishan, in summer 2006 and 2008. Without petals enclosing the obconical receptacle, floral temperature approximated to the ambient air temperature. A comparision was made of seed production of hand-pollinated D-1 flowers from which all petals were removed with that of flowers in which petals were intact. When petals were removed at dawn, the ambient temperature in Lantian, Hubei Province, south China (29°39′ N, 113°07′ E) was on average 7 °C higher than that in Mishan, Heilongjiang Province, north-east China (45°32′ N, 131°53′ E). The average July temperatures in Lantian and Mishan were 29·3 °C and 20–22 °C, respectively (Wang and Zhang, 2005).
Lotus ovules expand after fertilization, so seed set was evaluated as the ratio of expanded seeds to the number of ovules. Seed set among different treatments on lotus flowers was compared by one-way ANOVA using SPSS, version 13.0.
During the daytime refrigeration period, the temperature in the box interior fell immediately from about 30·1 to 19·2 °C when the ice bags were put in, and then rose again gradually. Another temperature decline occurred around noon from about 26·2 to 23·1 °C when melting ice bags were replaced with newly frozen ones, and then the temperature remained at around 23·0 °C (Fig. 1). Correspondingly, the temperature in unmanipulated control flowers during the daytime was relatively constant, increasing from 33·4 to 36·2 °C and then dropping back to 33·4 °C (Fig. 1). During nocturnal refrigeration, the temperature first dropped immediately from 30·0 to 15·8 °C when the ice bags were put in, and then increased gradually to 20·0 °C at the end of the night (Fig. 2). During the same period, unmanipulated control flowers maintained a stable temperature at around 32·0 °C (Fig. 2). Thus, our refrigeration treatment greatly reduced flower temperature.
Seed set was not significantly different between flowers that were hand-pollinated after refrigeration (treatment 1; mean ± s.e. = 0·85 ± 0·03, n = 13) or hand-pollinated without pre-refrigeration (treatment 2; 0·89 ± 0·01, n = 15; F1,27 = 2·12, P = 0·16), suggesting that stigma receptivity was not influenced by low floral temperature. Seed set of flowers pollinated with pollen treated with daytime refrigeration (0·93 ± 0·02, n = 13) and nocturnal refrigeration (0·97 ± 0·01, n =15) were not significantly different from those of flowers pollinated with normal pollen (0·94 ± 0·02, n = 13; F2,40 = 1·81, P = 0·18), indicating that low temperature in either period before pollen shedding did not affect pollen viability.
In two field populations, seed set of flowers from which petals were removed was significantly lower than that of intact flowers in the north-China population (t = 2·60, P = 0·01), but this difference was not observed in the south-China population (t = 0·52, P = 0·61; Fig. 3), suggesting that low ambient temperature reduced female reproductive success in lotus.
The temperature manipulations in the experimental plants indicated that daytime refrigeration immediately after pollination (treatment 3) significantly reduced seed set in lotus flowers (F3,75 = 497·94, P < 0·001), whereas there were no significant differences among flowers refrigerated during the night (treatment 4), flowers with petals removed (treatment 5) and control flowers (treatment 6; Fig. 4), indicating that low temperature significantly affected the fertilization process but not the post-fertilization events.
The results demonstrate that the stable temperature maintained in lotus flowers was an assurance for reproductive success when ambient temperatures are unfavourably low. The average air temperature on summer days is approx. 29·3 °C in south China with a maximum reaching 39·3 °C, but the average is lower in Mishan, north China, being only approx. 20–22 °C (Wang and Zhang, 2005). Seed set was not significantly different between flowers with petals removed and intact flowers in Wuhan and Lantian, Hubei Province, south China, where the ambient temperature in summer is high, but flowers with petals removed set fewer seeds than intact flowers in the cooler region of north China, confirming that low ambient temperature reduced reproductive success in lotus flowers (Figs 3, ,4).4). Kudo (1995) observed that the seed set of flowers of Adonis ramosa with petals removed was significantly lower than that of intact individuals in north Japan, indicating that elevation of temperature by heat absorption increased reproductive success by increasing fertilization success and seed development.
Compared to the post-fertilization process, the fertilization process was susceptible to low temperature in lotus. Low floral temperature during the fertilization period significantly reduced the seed set but low temperature during the phase of seed development did not, indicating that thermoregulation in lotus flowers was essential for the fertilization process rather than for the post-fertilization period (Fig. 4). Our experiments showed that low floral temperature did not affect stigma receptivity in D-1 flowers or pollen viability in undehisced anthers in the female phase. Therefore, the low seed set in flowers refrigerated during the fertilization process might be caused by the failure of some aspect of pollen performance, such as pollen germination, pollen-tube growth or fertilization in the pistils. The environment during stamen development has been shown to influence subsequent pollen quality and pollen performance in the recipient pistil in numerous plants (Schlichting, 1986; Young and Stanton, 1990; Delph et al., 1997; Aizen and Raffaele, 1998; Travers, 1999). The temperature in the environment of the stamens affected pollen quality and performance in the snow buttercup Ranunculus adoneus, where pollen from solar-tracking donor flowers exhibited a 32 % advantage in germination compared to that from stationary donor flowers (Galen and Stanton, 2003). The lack of any obvious effect of flower temperature on pollen quality in our study might be due to the short period of refrigeration or the fact that it was applied only during the late stages of pollen development. Given that the thermogenesis in lotus flowers begins 1–2 d before anthesis (Seymour and Schultze-Motel, 1998), the effect on pollen quality and performance of the resulting temperature elevation in the early stages of development needs further study. In adddition, we observed that anther dehiscence was delayed for about 5 h at low floral temperatures (see also Yates, 1993); thus, warming of the flowers may accelerate anther dehiscence in this aquatic plant.
Various functions have been proposed for floral thermogenesis: to protect flowers from freezing (Knutson, 1974); to enhance the emission of floral scents to attract pollinators (Fægri and van der Pijl, 1979; Meeuse and Raskin, 1988); to provide a direct energy reward for insect pollinators (Seymour and Schultze-Motel, 1997; Seymour et al., 2003); or to enable beetles to reach flight temperature and depart from the flower, carrying pollen away (Burquez et al., 1987). The lotus flower has a typical beetle-pollination syndrome, although diverse flower visitors including Coleoptera, Hymenoptera and Diptera (Sohmer and Sefton, 1978; Schneider and Buchanan, 1980) have been observed. Thermoregulation has been interpreted as rewarding endothermic beetles in addition to accelerating scent emission in lotus flowers (Schneider and Buchanan, 1980; Seymour and Schultze-Motel, 1996). Stigma receptivity in lotus flowers coincides with the peak of thermoregulation (Seymour and Blaylock, 2000), and thus the temperature in the floral chamber can greatly influence the post-pollination process, when viable pollen has been transported onto the receptive stigmas. We speculate that thermoregulation not only facilitates the arrival and departure of pollinators, increasing male reproductive success, but also that elevated temperature in female organs facilitates fertilization and seed development. Although the optimal range of temperature for fertilization success differs among species and cultivars of the same species, for example 20–25 °C in cherimoya (Rosell et al., 1999), 15–25 °C in mango (Sukhvibul et al., 2000) and 22–26 °C in papaya (Cohen et al., 1989), unfavourable temperatures, far from the optimum, usually lead to the failure of fertilization and to low seed set. Our experiments on hand-pollinated flowers showed that seed production was reduced in both refrigeration-treated flowers and wild flowers experiencing low temperature. Thus, this study has shown that production of a favourable temperature could also be crucial for post-pollination events in thermoregulatory plants.
The authors thank Wei-Dong Ke from Wuhan Institute of Vegetable Science, Hubei Province for his support during the work, Qi-Chao Wang from the Lotus Society of China for his encouragement and for suggesting this study, Zuo-Dong Li, En-Xing Zhou, Dong-Xu Li and Xiao-Xin Tang for their valuable help in the field, and Sarah Corbet for helpful comments on the manuscript. Grants from the National Science Foundation of China (no. 30825005) and the Ministry of Education of China (no. NCET-04–0668) to SQH supported this work.