Botanical Studies (2011) 52: 479-491.
REPRODUCTIVE BIOLOGY
Cryptic dioecy of Symplocos wikstroemiifolia Hayata (Symplocaceae) in Taiwan
Yu-Chen WANG and Jer-Ming HU*
Institute of Ecology and Evolutionary Biology, National Taiwan Uni-versity, Taipei 10617, Tai-wan
(Received November 25, 2010; Accepted March 24, 2011)
ABSTRACT. Symplocos wikstroemiifolia Hayata is one of the few morphologically androdioecious spe­cies in Symplocaceae. Although this species has been proposed as cryptically dioecious, little is known about its patterns of sexual expression in the field. We studied the breeding system and reproductive biology of S. wikstroemiifolia in Taiwan. Field investigations showed that anthers of most morphologically hermaphroditic flowers did not produce pollen grains; thus, this flower type should be considered female. Nearly all flowering individuals produced only male or female flowers. The sex ratios were slightly male-biased, but did not sig­nificantly deviate from 1:1, congruent with the proposed cryptic dioecy status. However, two individuals pro­duced flowers that may have been functionally hermaphroditic, suggesting a variation of sexual expression in S. wikstroemiifolia. There were about 12 stamens per male flower but only five staminodes per female flower. No locules and ovules existed in the male flowers, while two locules within the ovary of the female flowers con­tained four ovules each (two large, two small). Male individuals produced cymes and thyrses, whereas female individuals only produced cymes. Some of these morphological characteristics differed from those previously described. Anthesis of most male and female flowers began before dawn and lasted for one to two days. Nev­ertheless, anther dehiscence, which only occurred under dry conditions, was not restricted to a specific time period during the day. The mean natural fruit set was 6.7%, which is relatively low compared to those in other Symplocos or dioecious species. In conclusion, S. wikstroemiifolia exhibited a nearly complete gender dimor­phism, but with certain flexibility in sexual expression, and should thus be considered cryptically dioecious.
Keywords: Androdioecious; Breeding system; Cryptic dioecy; Gender dimorphism; Reproductive biology; Symplocos wikstroemiifolia.
INTRODUCTION
sterile and usually inaperturate pollen grains (Mayer and Charlesworth, 1991; Cane, 1993; Furness, 2007), such as in some species of Thalictrum (Kaplan and Mulcahy, 1971; Penny and Steven, 2009), Solanum (Anderson, 1979; Anderson and Symon, 1989; Knapp et al., 1998) and Vitis (Kevan et al., 1985; Kevan et al., 1988; Caporali et al., 2003). In the second type (type II), the functionally female flowers do not release pollen grains during anthe-sis, such as in Echinocereus coccineus (Hoffman, 1992), Withania aristata (Anderson et al., 2006) and in some spe-cies of Dombeya (Humeau et al., 1999). In the third type (type III), the functionally female flowers release viable pollen grains that can sire seeds; however, seedlings sired by functional females tend not to survive in natural popu-lations (Verdu et al., 2004). Thus, morphological form alone may not show its true reproductive function (Cruden and Lloyd, 1995; Sakai and Weller, 1999).
Androdioecy is a condition in which the male and her­maphroditic individuals coexist in a population, and it is extremely rare among plants (Darwin, 1877; Richards, 1997). Studies of many morphologically androdioecious plants often reveal that they are cryptically dioecious
Dioecy, a plant breeding system in which a population consists of distinct male and female individuals (Sakai and Weller, 1999), is found in fewer than 7% of angio-sperm species (Yampolsky and Yampolsky, 1922; Renner and Ricklefs, 1995; Richards, 1997). Dioecy however, is widespread among many plant families, and it has risen in­dependently from hermaphroditism through various evolu­tionary pathways many times (Bawa, 1980; Charlesworth, 2002). The degree of opposite sex organs reduction on a dioecious plant also varies among different taxa (Mitchell and Diggle, 2005). If the rudiments cause unisexual flow­ers to have a hermaphroditic appearance and make the dioecious condition hard to recognize, the breeding system is defined as cryptic dioecy (Mayer and Charlesworth, 1991).
Cryptic dioecy can be classified into three types. In the first type (type I), the functionally female flowers release
*Corresponding author: E-mail: jmhu@ntu.edu.tw.
480
Botanical Studies, Vol. 52, 2011
rather than truly androdioecious (Charlesworth, 1984; Pannell, 2002). However, such studies are very limited and require careful examination on the functionality of the flowers. One of the candidates showing androdioecy came from the genus Symplocos (Symplocaceae), found during our previous plant survey of Taiwan (Tseng et al., 2008). Symplocaceae, a family in the order Ericales, consists of about 300 tree or shrub species, occurring in humid tropi­cal or subtropical montane forests in eastern and southern Asia, America and Australia (Heywood et al., 2007; APG III, 2009). The family has been classified into two genera (genus Symplocos and Cordyloblaste) and several taxa above the species level (two subgenera, three sections, and two series in Symplocos; Fritsch et al., 2008).
The breeding system of the genus Symplocos is pre-dominately hermaphroditic (Heywood et al., 2007), but a few species show gender dimorphism (Aranha Filho et al., 2007, 2009a, b). Recently Aranha Filho et al. (2009a) doc-umented prevalent type I cryptic dioecy in Symplocos sec-tion Hopea, which is mainly distributed in South America (Fritsch et al., 2008; Aranha Filho et al., 2010). Many of these are morphologically androdioecious and at least 10 species examined by Aranha Filho et al. (2009a) produced sterile and malformed pollen grains on morphologically hermaphroditic flowers.
The only Asian representative of S. section Hopea is Symplocos wikstroemiifolia Hayata (Fritsch et al., 2008; Aranha Filho et al., 2009a). This species has been con­sidered "probably androdioecious" (Wu and Nooteboom, 1996). Nagamasu (1998) described S. wikstroemiifolia as having "bisexual (female)" flowers in the 2nd edition of the Flora of Taiwan. Wang (2000) also stated that the stamens of morphologically hermaphroditic flowers are "unde­veloped" in S. wikstroemiifolia. Although no detailed de­scription exists in the literature, this species seems likely cryptically dioecious.
In this paper, we provide evidence for the functional­ity of the flowers in S. wikstroemiifolia, and also basic descriptions of the reproductive biology of this species in Taiwan. Sexual expressions of every flowering individual from two natural populations in Taiwan were recorded for at least two consecutive years. Flower morphology, phe­nology and natural fruit set were also investigated.
ovary. There is one style on the morphologically hermaph­roditic flower but none on the male flower, and only one seed per drupe when mature.
Symplocos wikstroemiifolia is distributed in southern China, Indochina, the Malay Peninsula, and Taiwan (Na-gamasu, 1998). In Taiwan, the species is scattered across the middle and northern part of the island, from 700 m to 1,900 m altitude. They are not very common, and only locally abundant around Beichatian Mountain in Taiwan (Wang, 2000). Individuals start flowering from March to May, and set fruit from July to September (Wang, 2000).
Study sites
Field investigations were conducted at two sites where natural populations of S. wikstroemiifolia occur. One is located in the Chatianshan Nature Reserve, on the ridge around Beichatian Mountain (Figure 1). This area is one of the few places in Taiwan for Fagus hayatae (Fagaceae) (Liu and Su, 1972; Hsieh, 1989; Chiou et al., 1998; Lu et al., 1998). Beichatian Mountain is 1728 m high and S. wikstroemiifolia is particularly abundant at altitudes above 1,400 m on the surrounding ridge. The mean annual tem­perature is approximately 14.2°C (Chiou et al., 2004) and the mean annual precipitation, from 2005 to 2008, was 4,238 mm, as recorded by the Fushan Weather Station (about 5 km east of Beichatian Mountain and at 455 m al­titude).
The other study site was the mountainous area of Pin-glin District, New Taipei City (Figure 1). This site is along the roadside of Provincial Highway No. 9 near the "54.5 K" road sign located in a valley near the Sidu River at about 400 m altitude. The mean annual temperature, from
MATERIALS AND METHODS
Study species
Symplocos wikstroemiifolia Hayata is an evergreen shrub or tree with grayish or brownish bark. The leaves are simple, alternate, with an entire or sometimes a crenu-late margin (Wang, 2000). This species is morphologically androdioecious (Wu and Nooteboom, 1996; Nagamasu, 1998; Wang, 2000). The male flowers usually have more than 10 stamens, whereas the morphologically hermaph­roditic flowers have only five stamens. Both flower types are white, small and inconspicuous. They both have a five-lobed corolla and a nectary disk on the superior part of the
Figure 1. Study sites. Label A indicates the location of Beicha-tian Mountain. Label B indicates the second study site "Pinglin". The image is a magnification of a region from northern Taiwan. Scale bar = 25 km.
WANG and HU ― Cryptic dioecy of S. wikstroemiifolia
481
2005 to 2008, was 19.6°C, and the mean precipitation was 4,108 mm, based on the records in the Sidu Weather Sta­tion.
In the following text, we use the term "Beichatian Mountain" for the mountain ridge around Beichatian Mountain, and "Pinglin" for the second site. The distance between the two study sites is about 33 km.
slide. All grains in the droplet were counted under a light microscope. This procedure was repeated until all remain­ing solution was examined. After that, another 5 [L 70% ethanol was added to the tube and all previous counting procedures were repeated. When no more pollen grains were found in the 1 [L ethanol droplet after three succes­sive counts and either no or a negligible number of pollen grains were left in the PCR tube, the anther counts were completed. The sum of all numbers obtained from each counting procedure represented the total pollen grains per anther. The mean pollen grains per flower was calculated by multiplying mean pollen grains per anther by the mean stamens per flower.
To examine other flower characteristics, such as anther and ovary morphology, flower buds near anthesis were collected in the field and fixed in FAA. They were dehy­drated through a serial ethanol concentration gradient and transferred to Histo-Clear II® (Electron Microscopy Sci­ence, Pennsylvania, USA), then embedded in paraplast. Serial cross- and longitudinal sections were cut at 6-10 [im thickness using a rotary microtome, then were mounted onto slides that were coated with Mayer's adhesive. The slides were then stained with Safranin O and Fast Green, mounted with coverslips using Permount® (Fisher Scien­tific, New Jersey, USA), and observed under a light micro-scope.
Inflorescence morphology was observed in the field. Numbers of flower per inflorescence were counted from 101 inflorescences of six male individuals and 93 inflores­cences of eight morphologically hermaphroditic individu­als.
To compare the means in each measurement of different flower categories, students' t-tests or one-way ANOVAs were conducted. Flower characteristics were compared with either Welch's ANOVAs (Day and Quinn, 1989) or standard ANOVAs. Duncan's multiple range tests were ap­plied if ANOVAs showed a significant difference among different flower categories.
Patterns of sexual expression and sex ratio
From 2007 to 2009, a total of 190 S. wikstroemiifolia plants were labeled at Beichatian Mountain (188) and at Pinglin (2). During their 2007 and 2008 flowering sea­sons, individuals' floral morphologies were examined and recorded. We continued to monitor the flower produc­tion patterns in some individuals during 2009. Thirty or more flowers from each randomly selected individual's inflorescence were examined to verify the consistency of flower types. In small trees with fewer than 30 flowers, all flowers were observed. Flowers were inspected either in the field or in a laboratory under a stereomicroscope. The presence versus absence of pollen grains from the anthers of morphologically hermaphroditic flowers was carefully examined. The sex ratio, defined as the ratio of individuals with male flowers (male individuals or plants) to individu­als with morphologically hermaphroditic flowers (mor­phologically hermaphroditic individuals or plants) in a population, was calculated with data from 2007 and 2008. Chi-square tests with Yates' correction were conducted to determine whether the sex ratios significantly deviated from 1:1.
Flower and inflorescence morphology
The lengths of floral organs, i.e. sepal, petal, stamen, style, and ovary, were measured with the software ImageJ, which can analyze the pixel distance on a digital photo (Abramoff et al., 2004) (Table 2). Flowers were randomly collected in the field and fixed in FAA (formalin: glacial acetic acid: 95% ethanol: water = 2:1:10:7 in volume), and digitally photographed. The length and width of each floral organ were then measured and calculated with ImageJ on a computer. We did not measure all the floral organs within a single flower because some could not be fully flattened and stretched when they were photographed, which would have caused inaccurate measurements using ImageJ. The number of stamens per flower was recorded, and the num­ber of ovules per flower was observed from the dissected ovaries under a stereomicroscope.
To estimate pollen grain number per anther and flower, 10 flower buds near anthesis were collected from four male individuals and stored in 70% ethanol. From each flower bud, one to four indehiscent anthers were randomly chosen, and a total of 20 anthers were examined. An inde-hiscent anther was placed in a 0.2 mL PCR tube to which a droplet of 5 [iL 70% ethanol was added, and the anther was smashed with tweezers. After the anther released pollen grains, 1 [L of the suspension solution containing pollen grains was transferred and placed on a microscopic
Flowering phenology and natural fruit set
Symplocos wikstroemiifolia flowering phenology was observed at three different levels: the duration of the flow-ering season in one population, the flowering sequence within one inflorescence, and the flowering course during one day.
The duration of flowering was recorded from five male and five morphologically hermaphroditic plants in Be-ichatian Mountain in 2008. If individuals had fewer than about 100 inflorescences, the number of all inflorescences was counted. Otherwise, only several shoots were selected and the numbers of inflorescences among these shoots were counted. During the flowering season, the number of inflorescence with open flowers was monitored weekly. The number recorded was divided by the initial number of inflorescences per plant and expressed as a percentage. In Pinglin, in 2008, the same method was also applied to one male (Plant No. 115) and one morphologically hermaph-
482
Botanical Studies, Vol. 52, 2011
roditic plant (Plant No. 116). To investigate the flowering sequence within one inflorescence, flowering shoots that contained numerous inflorescences on Plant No. 115 and Plant No. 116 were observed for 12 consecutive days at Pinglin in 2009. The floral longevity and the flowering sequence were recorded with a digital camera. To under-stand the detailed flowering progression during one day, a continuous 24-hour observation was performed in Pinglin. During the observation periods, the flowering shoots on Plant No. 115 and 116 were photographed every one or two hours. The observation began at about 4:00 pm on March 22nd, 2009 and ended at about 6:00 pm on March 23rd, 2009.
The natural fruit set of S. wikstroemiifolia was recorded for 10 morphologically hermaphroditic plants in Beicha-tian Mountain in 2009. Numbers of flowers and flower buds near anthesis on each individual were counted dur­ing the flowering season. If the plants had fewer than 100
flowers, all the flowers and flower buds were counted. Otherwise, only part of the flowers and flower buds on se­lected flowering shoots and branches was counted. About two months later, the number of developing fruits was recorded, and the fruit set was calculated as the number of developing fruits divided by the number of flowers and flower buds.
RESULTS
Patterns of sexual expression and sex ratio
In the two study sites, from 2007 to 2009, 83 individu­als of S. wikstroemiifolia produced flowers during at least one flowering season. Only two flowering individuals were found in Pinglin, and the other 81 individuals were all located in Beichatian Mountain. Visual inspection of flowers showed that the anthers of male flowers all con­tained pollen grains (Figure 2A), whereas those of most
Figure 2. Flower and inflorescence morphology of S. wikstroemiifolia. (A) A male flower; (B) A female flower on one inflorescence; (C) A morphologically hermaphroditic flower from Plant No. 29. The anthers produced pollen grains; (D) A morphologically hermaph­roditic flower from Plant No. 189 (note the short style in the flower); (E) A cross-section of the proximal region of one ovary from a female flower. There were two locules in one ovary separated by one septum; (F) The ovules in one of the two locules (see the red circle). Two small ovules were above two large ovules; (G) An unbranched male cymose inflorescence. The apical flower opened first; (H) A branched male thyrsoid inflorescence; (I) A female cymose inflorescence.
WANG and HU ― Cryptic dioecy of S. wikstroemiifolia
483
morphologically hermaphroditic flowers were empty (Figure 2B). Serial sections of flower buds revealed the same pattern (Figures 3A, 3B and 3C). Morphologically hermaphroditic flowers contain normally developed ova­ries and ovules; therefore they can be considered having female-only function. To avoid confusion, we use the term "female flowers" to refer to the typical morphologically
hermaphroditic flowers that did not produce pollen grains. Thus, individuals with female flowers are described as "female individuals" or "female plants" in the following text. Two individuals (Plant No. 29 and 189) in Beichatian Mountain produced morphologically hermaphroditic flow­ers that bore pollen grains (Figures 2C and 2D). Plant No. 29 produced one inflorescence with a few female flowers in 2007, had no flower production in 2008, and produced one inflorescence with a few unusual hermaphroditic flow­ers in 2009. Anthers of the unusual hermaphroditic flower from Plant No. 29 in 2009 released pollen grains (Figure 2C). In contrast, Plant No. 189 produced unusual hermaph­roditic flowers both in 2008 and 2009. Not all anthers of unusual hermaphroditic flowers from Plant No. 189 bore pollen grains. In many flowers, some anthers were empty thus resembling those found in female flowers. The style lengths were shorter than those in female flowers, and highly variable (Figure 2D, Table 2). The stigma surface in Plant No. 29 and 189 was reduced compared to those in female plants (Figures 2B, 2C and 2D). Pollen grains from Plant No. 29 and 189 were similar in morphology to those from ordinary male plants. All of these pollen grains
Table 1. Sex ratios of flowering individuals of S. wikstroemii­folia in Beichatian Mountain.
Year
M
F
Sex Ratio
Chi-square and P value
2007 2008 Total
35 24 44
30 18
35
1:0.86 1:0.75 1:0.8
X2 = 0.25, P = 0.62 X2 = 0.6, P = 0.44
X = 0.81, P = 0.37
M= the number of male plants, F= the number of female plants. There are 17 male and 6 female plants that flowered in two of the three years; therefore, the figure in the bottom row represents the actual number of flowering individuals among the population during the study period.
Table 2. Flower morphology measurements and stamen number per flower of S. wikstroemiifolia.
Male
Female
Plant No. 29
Plant No. 189
F and P value
0.77 ± 0.018
0.56 ± 0.013
0.72 ± 0.025
0.7 ± 0.021
Sepal length (mm)
[1.122, 0.485]
[0.821, 0.382]
[0.963, 0.401]
[0.885, 0.454]
F = 34.98
(A) (55, 15)
(C) (69, 16)
(A,B) (24, 5)
(B) (24, 5)
2.9 ± 0.05
1.81 ± 0.028
2.48 ± 0.046
2.1 ± 0.052
F = 136.37
Petal length
(mm)
[3.756, 2.101]
[2.289, 1.141]
[2.848, 2.017]
[2.748, 1.557]
P*****
(A) (65, 15)
(D) (65, 16)
(B) (25, 5)
(C) (25, 5)
TD/-.+ n1 T'i
1.69 ± 0.034
1.17 ± 0.029
1.61 ± 0.031
1.29 ± 0.041
rp ― c n a ^
Petal width
(mm)
[2.337, 1.16]
[1.641, 0.646]
[1.935, 1.361]
[1.736, 0.853]
F = 59.41 P*****
(A) (64, 15)
(C) (65, 16)
(A) (25, 5)
(B) (25, 5)
Stamen length
2.31 ± 0.05
1.18 ± 0.037
2.07 ± 0.061
1.59 ± 0.05
F = 130.2
(mm)
[4.099, 1.167]
[1.883, 0.637]
[2.905, 1.138]
[1.985, 0.998]
P*****
(A) (138, 15)
(D) (63, 16)
(B) (34, 5)
(C) (29, 5)
1.15 ± 0.048
1.47 ± 0.059
1.46 ± 0.117
1.22 ± 0.11
tp——c A^
Ovary length (mm)
[1.521, 0.826]
[1.766, 0.951]
[1.85, 1.159]
[1.504, 0.989]
F = 6.41 P****
(B) (15, 15)
(A) (16, 16)
(A) (5, 5)
(B) (5, 5)
1.01 ± 0.038
1.09 ± 0.024
1.12 ± 0.03
1.08 ± 0.092
F = 1 49
kJ\ aiy W1U.L11
(mm)
[1.279, 0.712]
[1.242, 0.909]
[1.166, 1.015]
[1.293, 0.808]
P = 0.27
(15, 15)
(16, 16)
(5, 5)
(5, 5)
1.8 ± 0.088
2.29 ± 0.049
0.87 ± 0.251
IT' _ 1 fi C Q
Style length (mm)
N.A.
[2.38, 1.129]
[2.407, 2.134]
[1.666, 0.218]
F = 19.58
P*****
(B) (16, 16)
(A) (5, 5)
(C) (5, 5)
12.3 ± 0.39
4.6 ± 0.18
6.6 ± 0.4
6.2 ± 0.49
tp——c\n c o
Stamen number/
1
[18, 8]
[5, 3]
[8, 6]
[8, 5]
F = 9/.68 P*****
11U VV CI
(A) (40, 40)
(B) (16, 16)
(B) (5, 5)
(B) (5, 5)
Flower characteristics compared with Welch's ANOVAs are noted with The F and P values resulting from ANOVAs or Welch's ANOVAs are presented in the last column. The P value is denoted as **** and ***** when less than 0.005 and 0.0001, respec­tively. The mean with standard error (士 SE) for each measurement is listed on the first row of each cell. The maximum and the mini­mum values are given as [maximum, minimum] on the second row. Letters in the beginning of the third line represent significant differences between categories at the 0.05 level based on Duncan's multiple range test. The numbers of samples measured and flow­ers used in each measurement are noted in the third line as (samples, flowers). The styles in male flowers are absent; therefore, no length is indicated.
484
Botanical Studies, Vol. 52, 2011
were triporate and spheroidal, with small granules on the surface.
All other recorded flowering individuals produced only one type of flowers, either male or female. None of the 81 individuals changed their sexes throughout the observation period (2007-2009). In Pinglin, one flowering individual was male (Plant No. 115) and the other was female (Plant No. 116). Both of them flowered each year from 2007 to 2009. In Beichatian Mountain, 44 male and 35 female plants were found. Among these plants, 17 male and six
female plants flowered two of the three years from 2007 to 2009, and six males and nine females flowered every year. There were more flowering male than flowering female individuals in Beichatian Mountain in 2007 and 2008, as counted either separately or cumulatively, but the sex ra­tios, which do not include Plant No. 29 and 189, did not deviate significantly from 1:1 (Table 1).
Flower and inflorescence morphology
We measured flower morphologies for male flowers,
Table 3. The locule number per ovary and description of the ovule morphology in flowers from female plants and plants No. 29 and 189. N, the number of flowers examined.
Types and numbers of the locule and ovule in one flower
Female
No. 29
No. 189
(N = 16)
(N = 5)
(N = 5)
Two locules per ovary. Two large ovules and two small ovules in each locule
15
0
0
One locule per ovary. Two large ovules and two small ovules in each locule
1
5
2
One locule per ovary. One large ovule and one small ovule in each locule
0
0
1
No locule and ovule in the ovary
0
0
2
Figure 3. A male (A,C) and a female (B,D) flower bud of S. wikstroemiifolia. (A) A male flower bud in a longitudinal section; (B) A fe­male flower bud in a longitudinal section. The anthers were empty; (C) A close-up of one anther of a male flower bud in cross-section; (D) An ovary of a female flower in cross-section. AW, anther wall. Ov, ovule. Pe, petal. Se, sepal. Sl, style. Sp, septum. Arrows indicate the anthers in (A,B) and pollen grains in (C). Scale bars = 200 [im in (A, B, D); scale bar = 50 [im in (C).
WANG and HU ― Cryptic dioecy of S. wikstroemiifolia
485
female flowers, and morphologically hermaphroditic flow­ers from Plant No. 29 and 189. Table 2 shows the mean, standard error (SE), maximum, minimum, and sample size of each measurement in flower morphology and stamen number per flower. The morphological characteristics that were compared with Welch's ANOVAs are also presented in Table 2. Male flowers showed significantly longer and larger sepals, petals and stamens than female flowers or the two unusual plants (No. 29 and 189) (P < 0.0001), ex­cept that Plant No. 29 had a sepal length and petal width comparable to the males. Female flowers had significantly longer ovaries than the males and Plant No. 189 (F = 6.41, P < 0.005), but did not differ significantly from Plant No. 29. The width of the ovaries was not significantly different among the different flower types (F = 1.49, P = 0.27). The average style length was significantly different in female flowers, Plant No. 29 and 189 (F = 19.58, P < 0.0001). Morphologically hermaphroditic flowers from Plant No. 29 contained the longest mean style length compared to the others, whereas the mean style length was the short-est in flowers from Plant No. 189 (Table 2). The styles of some flowers from Plant No. 189 were highly reduced (Figure 2D).
The mean stamen number per flower significantly dif­fered among flower categories (Welch's ANOVA, F = 97.68, P < 0.0001) (Table 2). The value in male flowers ranged from 8 to 18, with a mean of 12.3 (SE = 0.39), which was significantly higher than that in the other flower types. Most male flowers contained fewer than 15 stamens (80% of 40 examined male flowers). The stamen (stamin-ode) number per female flower ranged from three to five, and most female flowers (69% of 16 examined female flowers) contained five stamens (staminodes). The range of stamen number per flower was six to eight and five to eight in Plant No. 29 and 189, respectively.
Figure 4. The flowering season duration of S. wikstroemiifolia in 2008. (A) Male plants in Beichatian Mountain. (B) Female plants in Beichatian Mountain. (C) Plants No. 115 and 116 in Pinglin. The y-axis indicates the percentage of inflorescences with open flowers.
Figure 5. Anthesis of male and female flowers in S. wikstroemiifolia. Photos were taken in Pinglin on March 23rd 2009. The time is indicated below each photo. (A) A male flower from Plant No. 115. The petals opened slightly between 1:34 am to 2:20 am, and the flower completely opened about 7:14 am. (B) Female flowers from Plant No. 116. The petals opened slightly before 3:11 am, and the flowers completely opened about 7:54 am.
486
Botanical Studies, Vol. 52, 2011
The number of pollen grains per anther in male flow­ers ranged from 181 to 1602, with a mean of 883 (SE = 101.12). The mean of pollen grains per male flower was estimated 10860.9 (SE = 1243.84).
In female flowers, we generally found two locules per ovary (Figures 2E and 3D), but occasionally a single loc-ule was found (sample from Plant No. 33) (Table 3). Each locule contained four pendulous and curved ovules, two large and two small, per locule (Figure 2F). The size of the large ovules was almost 1 mm long, whereas that of the small ovules was about half or one-third this size. In Plant No. 29, each morphologically hermaphroditic flower had only one locule, whereas in Plant No. 189, the five flow­ers examined showed different kinds of degeneration in the ovary. Two flowers had only one locule containing two large and two small ovules. The other single flower had only one large and one small ovule in one locule, whereas the remaining two flowers lacked both ovules and locules (Table 3). No ovules or locules existed in the ovaries of male flowers.
Male plants produced both unbranched (Figure 2G) and branched inflorescences (Figure 2H), whereas female plants produced only unbranched inflorescences (Figure 2I). The average flower number per inflorescence in male plants was 15.2 (SE = 0.89, range from 2 to 47), whereas that in female plants was 4.1 (SE = 0.21, range from 1 to 11). The difference was significant between the two sexes (t= 12.09, P < 0.0001).
After 2:00 am, the petals opened slowly and gradually, lasting about four hours or more. By 6:00 or 7:00 am, the petals had opened completely.
The flowering progress was similar for most female flowers. Their petals also gradually opened before dawn, and fully opened after sunrise. Both male and female flow-ers of S. wikstroemiifolia generally opened for one to two days, but a few flowers lasted for three to four days. The anther dehiscence in male flowers was mainly dependent on weather conditions. Anthers only dehisced under dry conditions; in humid weather, anthers would not dehisce even when the petals had fully opened for several hours. When fully opened, both male and female flowers pro-duced a small amount of nectar with their nectary disks.
The average natural fruit set was 6.7% (SE = 2.05%) and ranged from 0% to 17.8% among 10 female individu-als. We tagged an additional 144 flowers and flower buds in Plant No. 189, and recorded the natural fruit set in 2009. During the fruiting season only one developing fruit was found, so the fruit set was extremely low in this unusual individual (0.69%). The natural fruit set of Plant No. 29 was not investigated due to the scarcity of flowers on this individual in 2009.
DISCUSSION
The breeding system of Symplocos wikstroemi-ifolia
Flowering phenology and natural fruit set
Our study shows that S. wikstroemiifolia exhibits nearly complete gender dimorphism. Although female flowers display a hermaphroditic appearance, the anthers on these flowers do not produce any pollen grains. This result is compatible with Nagamasu's (1998) and Wang's (2000) descriptions, and reveals that this type of flower has only female reproductive function. During the period of field investigations, nearly all flowering individuals produced either only male or female flowers, and among these plants, none of them showed changes of sex. Symplocos wikstroemiifolia was thus confirmed as cryptically dioe­cious. However, two flowering individuals produced mor­phologically hermaphroditic flowers that produced pollen grains, styles and ovules; these may thus be functionally hermaphroditic (but in some flowers from Plant No. 189 the ovules were absent. See later discussion).
The sex ratios in the population were slightly male-biased, but did not significantly deviate from 1:1, which provided further evidence of cryptic dioecy (sensu Charlesworth, 1984; Verdu et al., 2004). In dioecious spe­cies, female plants usually invest more resources in repro­duction due to the production of fruits and seeds, whereas male plants lack this investment (Delph, 1999; Obeso, 2002; Wheelwright and Logan, 2004). Therefore, female plants may reproduce less frequently or reach reproductive maturity later than male plants (Delph, 1999; Queenbor-ough et al., 2007), which would result in a male-biased flowering sex ratio (Bullock and Bawa, 1981; Garcia and Antor, 1995; Obeso et al., 1998).
In Beichatian Mountain, the flowering season began in early April and ended in middle May, whereas in Pinglin it began in late March and ended in middle April (Figures 4A and 4B). The duration of flowering in each individual last­ed for two to three weeks. Among the observed flowering individuals in Beichatian Mountain, one male individual (Plant No. 186) began to flower earlier than the other male and female individuals, whereas in Pinglin, the male (Plant No. 115) and female (Plant No. 116) individual began to flower at approximately the same time (Figure 4C).
In most unbranched inflorescences from both male and female plants, the distal flower opened first, and so fit the definition of a cyme (Figures 2G and 2I). In most branched inflorescences from male plants, the lateral unit of inflores­cences was cymose because the distal flower opened first, but the main axis of the inflorescence was racemose such that the apical flower buds were still small when the flow­ers on the lateral unit of inflorescences opened. Branched inflorescences, therefore, fit the definition of thyrses (Fig­ure 2H). Despite the general trend, in a few unbranched inflorescences from both male and female plants the proxi­mal flower opened first.
Anthesis of most male and female flowers began before dawn (Figure 5). In most male flowers, the tips of the pet­als opened slightly in the beginning, and part of the an­thers were exposed. At this stage, the filaments were still covered by the petals and could not be seen from outside.
WANG and HU ― Cryptic dioecy of S. wikstroemiifolia
487
In other cryptically dioecious species in Symplocos section Hopea, the functionally female flowers produce inaperturate and sterile pollen grains (Aranha Filho et al., 2009a). These sterile pollen grains might be maintained as pollinator rewards (Cane, 1993; Kawagoe and Suzuki, 2004). However, female flowers of S. wikstroemiifolia did not bear any pollen grains, and although the morphologi­cally hermaphroditic flowers from Plant No. 29 and 189 produced pollen grains, the morphology of these grains was identical to those from the male flowers, triporate and spheroidal (Wang and Ou, 2000), not inaperturate (Aranha Filho et al., 2009a). According to Fritsch et al. (2008), S. wikstroemiifolia is sister to S. tinctoria, and the next two related lineages are S. longipes and S. guadeloupensis. These species, except S. wikstroemiifolia, are all hermaph­roditic. Other cryptically dioecious species within this section form a monophyletic group, which is sister to S. guadeloupensis. Therefore dioecy is probably an autapo-morphy for S. wikstroemiifolia. This interpretation can ex­plain why only S. wikstroemiifolia displays type II cryptic dioecy in S. section Hopea.
at the base of the unbranched inflorescences sometimes opened prior to the others, the apical flower buds at that time were well-developed and near anthesis. We suggest that these unbranched inflorescences could be derived from thyrsoid inflorescences with some reductions of the lateral units. To verify this hypothesis, it will be necessary to observe the development of the inflorescence buds.
The thyrsoid inflorescences in male plants that allow a higher number of flowers per inflorescence, together with the larger petals size of male flowers (Table 2), provide a larger floral display that may attract more pollinators than do the female plants. Males tend to be limited by mating opportunities, whereas females tend to be limited by the resources necessary for producing offspring (Bateman, 1948; Bell, 1985; Arnold, 1994). Male plants may require a higher pollinator visitation rate compared to female plants. Some studies indicate that pollinators tend to visit plants with a larger number or size of flowers or inflores­cences more frequently (Conner and Rush, 1996; Vaugh-ton and Ramsey, 1998; Ohashi and Yahara, 2001). Male individuals may therefore be more advantageous when showing larger floral display (Bell, 1985). Another expla­nation for this sexual dimorphism is that because female individuals generally incur a higher cost of reproduction, they cannot produce as many flowers as males (Bullock and Bawa, 1981; Armstrong and Irvine, 1989; Carr, 1991).
Flower and inflorescence morphology
Several flower and inflorescence characteristics of S. wikstroemiifolia examined in this study differed from those in previous descriptions. The first difference was the stamen number per male flower. Symplocos wikstroemiifo­lia was reported to have 15 to 20 stamens in a male flower (Wu and Nooteboom, 1996; Nagamasu, 1998; Wang, 2000), but in our measurements, the range was 8 to 18, with most male flowers (80%) having fewer than 15 sta­mens.
The second difference was the locule number per ovary in the female flowers. Although the ovaries of female flow­ers in S. wikstroemiifolia were described as having three locules (Nagamasu, 1998; Wang, 2000), i.e., the same number as the other species in S. section Hopea (Fritsch et al., 2008), we generally found only two locules and in a few female flowers, only one. The ovule number per loc-ule varies from two to four in Symplocaceae (Heywood et al., 2007) and was previously described as four in S. wik-stroemiifolia (Nagamasu, 1998; Wang, 2000). Although we indeed found four ovules per locule, two of the four ovules were significantly larger than the others. To our knowledge, the ovule dimorphism has not previously been reported in Symplocaceae. We suggest that the two smaller ovules represent a type of early abortion, and fertilization may only occur in the two larger ovules.
Finally, the inflorescence morphology was also incon-gruent with published descriptions. Nagamasu (1998) con­sidered the inflorescences to be panicles in male plants and racemes in female plants, respectively. In contrast, Wang (2000) described the inflorescence as spikes. However, we describe the inflorescences as cymes or thyrses in male plants and cymes in female plants, based on the flowering sequence within one inflorescence. Although the flowers
Flowering phenology and natural fruit set
In most dioecious species, the beginning of the flower­ing season is earlier in male than in female plants (Lloyd and Webb, 1977). In Beichatian Mountain, one male indi­vidual of S. wikstroemiifolia began to flower prior to the others observed (Figures 4A and 4B). In the first week of April, 2008 we found only male but not female flowers in this site. This suggests that male plants of S. wikstroemii-folia tend to flower earlier than females in the population. The flowering season began earlier in Pinglin than in Be-ichatian Mountain. The lower mean annual temperature in Beichatian Mountain may have caused the delay of floral initiation in S. wikstroemiifolia.
The natural fruit set of S. wikstroemiifolia in Beichatian Mountain was quite low (6.7%) compared to the value for other dioecious species. Bawa and Opler (1975) described the mean fruit set in some tropical forest dioecious trees as about 26.3%, and Sutherland and Delph (1984) even reported an average of 73.8% fruit set in dioecious spe­cies. The fruit set of S. wikstroemiifolia was also relatively low when compared with other Symplocos species. Wes-selingh et al. (1999) recorded a Symplocos species with an average of 26.9% fruit set in a tropical montane forest in Costa Rica. Devy and Davidar (2006) reported a mean fruit set of 46% for S. wynadense in a wet evergreen forest in India. These comparisons raise the question of what fac­tors cause the low fruit set in S. wikstroemiifolia. Further pollinator activity and pollen supplementation studies are essential to explain these factors.
488
Botanical Studies, Vol. 52, 2011
Unusual hermaphroditic flowers
studies. More extensive research focusing on flower devel­opment and pollination ecology is necessary to understand the evolution of dioecy in Symplocaceae.
Although S. wikstroemiifolia displayed nearly complete dioecy, the population still showed slight gender vari­ability. Among all the flowering individuals, we found two plants that produced morphologically hermaphroditic flowers that contained pollen grains. Several morphologi­cal traits of these unusual hermaphroditic flowers, such as perianth size and stamen number per flower, were between the range of male and female flowers and thus did not sim­ply contain the organ of the opposite sex in the flowers. It seems that the recovery of hermaphroditic reproductive function may alter the development of other flower charac­teristics, not only the fertile, but the sterile parts.
Furthermore, our observations suggest that the male and female reproductive abilities in these unusual her­maphroditic flowers may be limited, compared to ordinary male and female flowers. Fewer stamens were present in these than in male flowers. Moreover, not all anthers con­tained pollen grains in flowers from Plant No. 189. These flowers may produce fewer pollen grains. Significantly fewer ovules were also present in the morphologically hermaphroditic flowers, some had no ovules at all. Their style length is highly variable, and their stigmatic surface area is smaller than on those of female flowers. These fac­tors may result in low female reproductive function, as revealed by the extremely low natural fruit set in Plant No.
189 (0.69%).
The origin and reproductive role of the unusual her­maphroditic individuals of S. wikstroemiifolia remain unclear. The sexual expression pattern in Plant No. 29 in­dicated that a few female individuals of S. wikstroemiifolia might recover partial male reproductive function, since in 2007, this plant produced relatively fewer female flowers. Other flowering individuals of S. wikstroemiifolia did not change their sex during the observation period, thus only a few individuals in the S. wikstroemiifolia population may have the potential to become hermaphroditic. The occa­sional existence of hermaphrodites has been documented in many dioecious plant species (Lloyd and Bawa, 1984). When pollen transfers are insufficient, the occasional her­maphrodites may produce seeds through selfing (Baker and Cox, 1984), and thus allow dioecious species to per­sist. However, no evidence suggests that these unusual hermaphroditic flowers of S. wikstroemiifolia are capable of self-fertilization. Future studies should investigate such male and female reproductive functions as pollen fertility and stigma receptivity on these flowers to examine their potential for self-fertilization. Sexual expression studies of longer duration will also help elucidate the reproductive roles in these plants.
Our research clearly demonstrates that S. wikstroemiifo-lia is cryptically dioecious, but a few individuals showed incomplete gender dimorphism differentiation. Several flower and inflorescence characteristics of S. wikstro-emiifolia differ from those of previous descriptions. The level of natural fruit set for S. wikstroemiifolia was lower compared to those in other dioecious or Symplocos species
Acknowledgments. We thank Drs. Sheng-You Lu and Chih-Chiang Wang for providing information on the dis­tribution of Symplocos wikstroemiifolia in Taiwan; Mr. Juinn-Yi Huang for his helpful advice on the fieldwork; and Drs. Chih-Hua Tsou, Shau-Ting Chiu, Chun-Neng Wang and Kuo-Fang Chung, Peter W. Fritsch and one anonymous reviewer for their valuable comments on this work. Our gratitude is also due to Dr. Yu-Pin Cheng for his kind support and assistance.
LITERATURE CITED
Abramoff, M.D., P.J. Magalhaes, and S.J. Ram. 2004. Image processing with ImageJ. Biophotonics Int. 11: 36-42.
Anderson, G.J. 1979. Dioecious Solanum species of hermaphro­ditic origin is an example of a broad convergence. Nature 282: 836-838.
Anderson, G.J., G. Bernardello, M.R. Opel, A. Santos-Guerra, and M. Anderson. 2006. Reproductive biology of the dioe­cious Canary Islands endemic Withania aristata (Solan-aceae). Am. J. Bot. 93: 1295-1305.
Anderson, G.J. and D.E. Symon. 1989. Functional dioecy and andromonoecy in Solanum. Evolution 43: 204-219.
APG III. 2009. An update of the Angiosperm Phylogeny Group classification for the orders and families of flowering plants: APGIII. Bot. J. Linn. Soc. 161: 105-121.
Aranha Filho, J.L.M., P.W. Fritsch, F. Almeda, and A.B. Martins. 2007. Symplocos saxatilis (Symplocaceae), a new dioecious species from Serra do Cipo, Minas Gerais, Brazil. Brittonia 59: 233-237.
Aranha Filho, J.L.M., P.W. Fritsch, F. Almeda, and A.B. Martins. 2009a. Cryptic dioecy is widespread in South American species of Symplocos section Barberina (Symplocaceae). Plant Syst. Evol. 277: 99-104.
Aranha Filho, J.L.M., P.W. Fritsch, F. Almeda, and A.B. Martins. 2009b. Two new dioecious species of Symplocos (Symplo-caceae) from Southern Brazil. Novon 19: 1-6.
Aranha Filho, J.L.M., P.W. Fritsch, F. Almeda, and A.B. Martins. 2010. (1966) Proposal to reject the name Barberina hirsuta (Symplocos hirsuta) (Symplocaceae). Taxon 59: 1287-1288.
Armstrong, J.E. and A.K. Irvine. 1989. Flowering, sex ratios, pollen-ovule ratios, fruit set, and reproductive effort of a dioecious tree, Myristica insipida (Myristicaceae), in two different rain forest communities. Am. J. Bot. 76: 74-85.
Arnold, S.J. 1994. Bateman's principles and the measurement of sexual selection in plants and animals. Am. Nat. 144: S126-
S149.
Baker, H.G. and P.A. Cox. 1984. Further thoughts on dioecism and islands. Ann. Missouri Bot. Gard. 71: 244-253.
Bateman, A.J. 1948. Intra-sexual selection in Drosophila. Hered­ity 2: 349-368.
WANG and HU ― Cryptic dioecy of S. wikstroemiifolia
489
Bawa, K.S. 1980. Evolution of dioecy in flowering plants.
Annu. Rev. Ecol. Syst. 11: 15-39.
Bawa, K.S. and P.A. Opler. 1975. Dioecism in tropical forest
trees. Evolution 29: 167-179.
Bell, G. 1985. On the function of flowers. Proc. R. Soc. Lond.
B. 224: 223-265.
Bullock, S.H. and K.S. Bawa. 1981. Sexual dimorphism and the annual flowering pattern in Jacaratia dolichaula (D. Smith) Woodson (Caricaceae) in a Costa Rican rain forest. Ecology
62: 1494-1504.
Cane, J.H. 1993. Reproductive role of sterile pollen in crypti­cally dioecious species of flowering plants. Curr. Sci. 65:
223-225.
Caporali, E., A. Spada, G. Marziani, O. Failla, and A. Scienza. 2003. The arrest of development of abortive reproductive organs in the unisexual flower of Vitis vinifera ssp. silves-
tris. Sex. Plant Reprod. 15: 291-300.
Carr, D.E. 1991. Sexual dimorphism and fruit production in a di­oecious understory tree, Ilex opaca Ait. Oecologia 85: 381­388.
Charlesworth, D. 1984. Androdioecy and the evolution of di-
oecy. Biol. J. Linn. Soc. 22: 333-348.
Charlesworth, D. 2002. Plant sex determination and sex chromo­somes. Heredity 88: 94-101.
Chiou, C.A., K.C. Lu, and C.H. Ou. 1998. Studies on the veg­etation of Chatienshan Nature Reserve. J. Exp. For. Natl. Chung Hsing Univ. 20: 57-80.
Chiou, C.R., Y.C. Liang, Y.J. Lai, and M.Y. Huang. 2004. A
study of delineation and application of the climatic zones in Taiwan. J. Taiwan Geogr. Inf. Sci. 1: 41-62.
Conner, J.K. and S. Rush. 1996. Effects of flower size and number on pollinator visitation to wild radish, Raphanus raphanistrum. Oecologia 105: 509-516.
Cruden, R.W. and R.M. Lloyd. 1995. Embryophytes have equiv­alent sexual phenotypes and breeding systems: why not a common terminology to describe them? Am. J. Bot.
82: 816-825.
Day, R.W. and G.P. Quinn. 1989. Comparisons of treatments after an analysis of variance in ecology. Ecol. Monogr. 59:
433-463.
Darwin, C. 1877. The Different Forms of Flowers on Plants of the Same Species, D. Appleton and Company, New York,
352 pp.
Delph, L.F. 1999. Sexual dimorphism in life history. In M.A. Geber, T.E. Dawson, and L.F. Delph (eds.), Gender and Sexual Dimorphism in Flowering Plants. Springer-Verlag, Berlin, Heidelberg, pp. 149-173.
Devy, M.S. and P. Davidar. 2006. Breeding systems and pollina­tion modes of understorey shrubs in a medium elevation wet evergreen forest, southern Western Ghats,
India. Curr. Sci. 90: 838-842. Fritsch, P.W., L.M. Kelly, Y. Wang, F. Almeda, and R. Kriebel.
2008. Revised infrafamilial classification of Symplocaceae based on phylogenetic data from DNA sequences and mor-
phology. Taxon 57: 823-852.
Furness, C.A. 2007. Why does some pollen lack apertures? A review of inaperturate pollen in eudicots. Bot. J. Linn. Soc.
155: 29-48.
Garcia, M.B. and R.J. Antor. 1995. Sex ratio and sexual dimor­phism in the dioecious Borderea pyrenaica (Dioscoreaceae).
Oecologia 101: 59-67.
Heywood, V.H., R.K. Brummitt, A. Culham, and O. Seberg. 2007. Flowering Plant Families of the World. Firefly Books Ltd., New York, Ontario, 424 pp.
Hoffman, M.T. 1992. Functional dioecy in Echinocereus coc-cineus (Cactaceae): breeding system, sex ratios, and geo­graphic range of floral dimorphism. Am. J. Bot. 79: 1382­1388.
Hsieh, C.F. 1989. Structure and floristic composition of the beech forest in Taiwan. Taiwania 34: 28-44.
Humeau, L., T. Pailler, and J.D. Thompson. 1999. Cryptic dioecy and leaky dioecy in endemic species of Dombeya (Sterculi-aceae) on La Reunion. Am. J. Bot. 86: 1437-1447.
Kaplan, S.M. and D.L. Mulcahy. 1971. Mode of pollination and floral sexuality in Thalictrum. Evolution 25: 659-668.
Kawagoe, T. and N. Suzuki. 2004. Cryptic dioecy in Actinidia polygama: a test of the pollinator attraction hypothesis.
Can. J. Bot. 82: 214-218.
Kevan, P.G., D.C.A. Blades, U. Posluszny, and J.D. Ambrose. 1988. Pollen dimorphism and dioecy in Vitis aestivalis. Vi-
tis 27: 143-146.
Kevan, P.G., R.W. Longair, and R.M. Gadawski. 1985. Dioecy and pollen dimorphism in Vitis riparia (Vitaceae). Can. J.
Bot. 63: 2263-2267.
Knapp, S., V. Persson, and S. Blackmore. 1998. Pollen morphol­ogy and functional dioecy in Solanum (Solanaceae). Plant
Syst. Evol. 210: 113-139.
Liu, T.R. and H.J. Su. 1972. Synecological survey on the sum­mer-green forest of North Chia-tien Mountains. Sci. Annu. Taiwan Prov. Mus. 15: 1-16.
Lloyd, D.G. and K.S. Bawa. 1984. Modification of the gender of seed plants in varying conditions. Evol. Biol. 17: 255-338.
Lloyd, D.G. and C.J. Webb. 1977. Secondary sex characters in
plants. Bot. Rev. 43: 177-216.
Lu, K.C., C.H. Ou, and C.A. Chiou. 1998. Studies on the vegeta­tion of Chatienshan Nature Reserve (II): the population con­stitution of Taiwan beech. J. Exp. For. Natl. Chung
Hsing Univ. 20: 79-91.
Mayer, S.S. and D. Charlesworth. 1991. Cryptic dioecy in flow­ering plants. Trends Ecol. Evol. 6: 320-325.
Mitchell, C.H. and P.K. Diggle. 2005. The evolution of uni­sexual flowers: morphological and functional convergence results from diverse developmental transitions. Am. J. Bot.
92: 1068-1076.
Nagamasu, H. 1998. Symplocaceae. In T.C. Huang and Editorial Committee of the Flora of Taiwan (eds.), Flora of Taiwan, 2nd edn., Vol. 4. Editorial Committee of the Flora of Tai­wan, Taipei, pp. 101-127.
490
Botanical Studies, Vol. 52, 2011
Obeso, J.R. 2002. The costs of reproduction in plants. New
Phy-tol. 155: 321-348.
Obeso, J.R., M. Alvarez-Santullano, and R. Retuerto. 1998. Sex ratios, size distributions, and sexual dimorphism in the dioe­cious tree Ilex aquifolium (Aquifoliaceae). Am. J. Bot. 85:
1602-1608.
Ohashi, K. and T. Yahara. 2001. Behavioral responses of pollina­tors to variation in floral display size and their influences on the evolution of floral traits. In L. Chittka and J.D. Thomson (eds.), Cognitive Ecology of Pollination: Animal Behavior and Floral Evolution. Cambridge University Press, Cam­bridge, pp. 274-296.
Pannell, J.R. 2002. The evolution and maintenance of androdi-
oecy. Annu. Rev. Ecol. Syst. 33: 397-425.
Penny, R.H. and J.C. Steven. 2009. Sexual dimorphism in pollen grain size in cryptically dioecious Thalictrum macrostylum.
Plant Syst. Evol. 279: 11-19.
Queenborough, S.A., D.F.R.P. Burslem, N.C. Garwood, and R. Valencia. 2007. Determinants of biased sex ratios and inter-sex costs of reproduction in dioecious tropical forest trees.
Am. J. Bot. 94: 67-78.
Renner, S.S. and R.E. Ricklefs. 1995. Dioecy and its correlates in the flowering plants. Am. J. Bot. 82: 596-606.
Richards, A.J. 1997. Plant Breeding Systems, 2nd edn. Chapman & Hall, London, 529 pp.
Sakai, A.K. and S.G. Weller. 1999. Gender and sexual dimor­phism in flowering plants: a review of terminology, biogeo-graphic patterns, ecological correlates, and phylogenetic approaches. In M.A. Geber, T.E. Dawson, and L.F. Delph (eds.), Gender and Sexual Dimorphism in Flowering Plants. Springer-Verlag, Berlin, Heidelberg, pp. 1-31.
Sutherland, S. and L.F. Delph. 1984. On the importance of male fitness in plants: patterns of fruit-set. Ecology 65: 1093­1104.
Tseng, Y.H., C.F. Hsieh, and J.M. Hu. 2008. Incidences and eco­logical correlates of dioecious angiosperms in Taiwan
and its outlying Orchid Island. Bot. Stud. 49: 261-276.
Vaughton, G. and M. Ramsey. 1998. Floral display, pollinator visitation and reproductive success in the dioecious peren­nial herb Wurmbea dioica (Liliaceae). Oecologia 115: 93­101.
Verdu, M., A.I. Montilla, and J.R. Pannell. 2004. Paternal effects on functional gender account for cryptic dioecy in a peren­nial plant. Proc. R. Soc. Lond. B. 271: 2017-2023.
Wang, C.C. 2000. A taxonomic study of the Symplocaceae of Taiwan. Ph.D. thesis. National Chung Hsing University, Taichung, 358 pp.
Wang, C.C. and C.H. Ou. 2000. Study on the pollen morphology of Symplocos (Symplocaceae) from Taiwan. Q.
J. For. Res. 22: 21-36.
Wesselingh, R.A., M. Witteveldt, J. Morissette, and H.C.M. den Nijs. 1999. Reproductive ecology of understory species in a tropical montane forest in Costa Rica. Biotropica 31: 637­645.
Wheelwright, N.T. and B.A. Logan. 2004. Previous-year repro­duction reduces photosynthetic capacity and slows lifetime growth in females of a neotropical tree. Proc. Natl. Acad.
Sci. USA 101: 8051-8055.
Wu, R.F. and H.P. Nooteboom. 1996. Symplocaceae. In Z.Y.
Wu and P.H. Raven (eds.), Flora of China, Vol. 15. Sci­ence Press, Beijing & Missouri Botanical Garden Press, St. Louis, pp. 235-252.
Yampolsky, C. and H. Yampolsky. 1922. Distribution of sex forms in the Phanerogamic flora. Bibliotheca Genet. 3:
1-62.
WANG and HU ― Cryptic dioecy of S. wikstroemiifolia
491
台灣產荛花葉灰木之隱性雌雄異株之研究
王有禎 胡哲明
國立台灣大學生態學與演化生物學研究所
薨花葉灰木(Symplocos wikstroemiifolia Hayata)是灰木科（Symplocaceae)中少數外觀形態為雄花/兩
性異株花（androdioecy)的物種。雖然過往的研究指出薨花葉灰木可能是隱性雌雄異株（cryptic dioecy),
但是對該物種在野外生育地的性別表現狀況並不清楚。我們針對台灣產的薨花葉灰木進行生殖系統
(breeding system)與繁殖生物學（reproductive biology)的研究。生殖系統的調查結果顯示，來自外觀兩性
花上的雄蕊並不具花粉，因此這種型態的花朵應被視為雌花。幾乎所有的開花植株在調查期間只開雄花
或雌花，族群中雖然記錄到較多的雄株，但是性別比例並沒有明顯偏離1:1 。這些結果符合預想中隱性
雌雄異株的特徵。然而，有兩株個體產生了一些可能具有兩性生殖功能的花朵，這顯示少數植株在性別
表現上仍具備可塑性。花部外形測量顯示雄花平均具有約12枚雄蕊但雌花只具有5枚假雄蕊。每朵雄
花並不具有腔室或胚珠，而每朵雌花其子房具有兩室，而每室中具有4枚胚珠，且胚珠依照大小可以分
成兩型。雄株會產生不分支的聚饊花序（cymes))與密錐花序（thyrses))，而雌株只會產生不分支的聚饊花
序。以上這些外形特徵有些與過往文獻中的敘述並不一致。開花物候觀察顯示大部分的雄花和雌花在凌
晨開放花苞，且單花壽命大約是一至兩天，然而，雄蕊只在乾燥的環境下開裂，因此開裂的時間並不固
定。族群中的天然結實率為6.7% ，比起其他灰木屬或是雌雄異株的物種，薨花葉灰木具有較低的天然
結實率。總而言之，薨花葉灰木的性別表現幾乎完全分化為雄性與雌性兩種型態（gender dimorphism),
但是還是有一些小變異存在於族群中，因此這個物種應被視為隱性雌雄異株。
關鍵詞：雄花/兩性異株花；生殖系統；隱性雌雄異株；性別表現分化；生殖生物學；薨花葉灰木。