Patterns of plant sexual systems in subtropical evergreen broad-leaved forests in Ailao Mountains, SW China

Abstract

Aims

Sexual systems influence many components of the ecology and evolution of plant populations and have rarely been documented in subtropical evergreen broad-leaved forests (SEBLF). Here we report frequency distribution and ecological correlates of plant sexual systems in SEBLF at Ailao Mountains, and compare our results with that of tropical and cool temperate forests.

Methods

We examine the sexual systems of 703 species of woody angiosperms belonging to 103 families and 296 genera based on a comprehensive survey of SEBLF at Ailao Mountains Natural Reserve. Information of plant sexual systems and ecological traits were mainly based on published literatures and specimens as well as our field observations. The associations between plant sexual system and ecological traits were assessed with chi-square tests.

Important findings

Among these species, 60.2% were hermaphroditic, 15.8% were monoecious and 24.0% were dioecious. The percentage of dioecious sexual system among tree species (22.2%) in SEBLF was comparable to those of tropical tree floras, but much higher than those of temperate forests at North America. The percentage of monoecious sexual system among tree species (30.1%) in SEBLF was higher than that of tropical tree floras, but much lower than those of temperate forests at North America. Monoecy was significantly associated with the tree growth form and was relatively common in temperate genera. Dioecy was significantly associated with fleshy fruits and monoecy was significantly associated with dry fruit type in SEBLF. The high percentage of diclinous sexual systems (monoecy and dioecy) in SEBLF may be associated with the origin of the flora and the prevalence of relatively small inconspicuous flowers of the forests in the Oriental Region.

INTRODUCTION

Plant reproduction needs vectors (animals, wind or water) to transfer pollen and spread seeds due to their immobility. This reliance on vectors promotes the evolution of plant reproductive traits to manipulate agents, mainly animals, which will disperse pollen and seeds (Barrett and Harder 1996). Reproductive traits also play an important role in determining plant fitness, stand dynamics, genetic diversity and species interactions both at the population and community levels (Chazdon et al. 2003).

Sexual systems have been one of the most important reproductive traits for study in surveys of tropical forests because of the key role they play in influencing ecological and evolutionary components of plant populations (Bawa et al. 1985a; Bullock 1985; Carpenter et al. 2003; Chazdon et al. 2003; Flores and Schemske 1984; Gross 2005; Ibarra-Maríquez and Oyama 1992; Matallana et al. 2005). Significant effort has been spent in documenting correlations between plant sexual systems and morphological and ecological traits (reviewed in Geber et al. 1999; Vamosi et al. 2003). For example, Renner and Ricklefs (1995) found that the traits most consistently associated with dioecy at the family level are monoecy, wind or water pollination and the climbing growth form. Ecological correlates may provide the basis for testable hypotheses, although they do not reveal evolutionary mechanisms or selective pathways (Flores and Schemske 1984).

Most community-level surveys of plant reproductive traits were conducted in tropical regions (e.g. Bawa et al. 1985a, 1985b; Bullock 1985; Chazdon et al. 2003; Chen and Li 2008; Flores and Schemske 1984; Gross 2005; Ibarra-Maríquez and Oyama 1992; Matallana et al. 2005). This may be because of the general interests of plant ecologists attracted by high species diversity and complicated species interactions of tropical forests. However, patterns of plant reproductive traits in another important global vegetation type that contributes to both biodiversity and human welfare, subtropical evergreen broad-leaved forests (SEBLF), have been received much less attention than it should be.

SEBLF extensively distributed in East Asia, west coast of Atlantic of North America, Mediterranean coast of Europe, as well as local habitats of Oceania, South America and Africa (Ding and Song 2004; Song et al. 2005). SEBLF is well developed in East Asia and distributed mainly in southern China, Japan and southern Korea Peninsular (Li 1997). One important characteristics of SEBLF in East Asia is that the upper part of tree layer is mainly composed of evergreen broad-leaved trees in Fagaceae, Theaceae, Magnoliaceae and Lauraceae (Wu 1987). In China, SEBLF distributes between 24° N and 32° N and 99° E and 123° E and previously covered around 25% of the area of the country but declined rapidly due to human exploitation and destruction. Ailao Mountains (ca. 23°35′–24°44′ N, 100°54′–101°30′ E) are located in the central Yunnan Province and are at the geographical junction of Yunnan-Guizhou Plateau, Hengduan Mountains and Tibet Plateau (Fig. 1). The SEBLF is distributed widely and continuously in the upper part of the mountain, approximately 2 200–2 700 m above sea level (asl). In the inventory of the flora, 1 486 vascular plant species belonging to 720 genera and 207 families were recorded (Xu and Jiang 1988). Based on the analysis of 622 genera, 366 genera (59%) are tropical genera, 143 (23%) are temperate genera, 41 (7%) have world-wide distribution and 9 of them (1%) are endemic to China (Xu and Jiang 1988).

Patterns of plant sexual systems in subtropical regions have been rarely studied. Here, we addressed three principal objectives based on a survey of 703 woody angiosperm species in SEBLF at Ailao Mountains (i) to determine the frequency distribution of hermaphroditic, monoecious and dioecious sexual systems; (ii) to quantify the association between sexual systems and several morphological and ecological traits and (iii) to compare the frequency distribution of sexual systems among floristic components and with that of tropical communities/floras and temperate forests of North America.

MATERIALS AND METHODS

Study site

This survey was conducted at the Ailao Mountains Natural Reserve, Yunnan Province, SW China. Ailao Mountains are dominated by subtropical monsoon climate and in the transitional zone of middle subtropic and south subtropic climate in Yunnan Province (Zhang 1983). SEBLF is the main vegetation type in the reserve and has an area of ∼33 600 ha. The climate of Ailao Mountains has distinct dry and wet seasons. According to the meteorological data from Ailaoshan Forest Ecosystem Research Station (2 450 m asl), the annual precipitation is 1 931 mm, with over 85% of rainfall in the wet season (May–October), and the annual mean temperature is 11.3°C, with the hottest month 15.6°C and the coldest month 5.4°C (Qiu 1998).

Data collection

The data set used for this study was based on a species list compiled from the Comprehensive Report on Scientific Exploration of the Ailao Mountains Natural Reserve (Xu and Jiang 1988). The original species list included 1 344 angiosperm species. We excluded herbs, non-indigenous species and cultivated species.

Information of plant reproductive characteristics was mainly based on published literatures (Delectis Florae Reipublicae Popularis Sinicae Agendae Academiae Sinicae Edita 1958–2002; Institutum Botanicum Kunmingense Academiae Sinicae Edita 1975–2003; Flora of China Editorial Committee 1994–2005) and specimens (Herbaria KUN and HITBC) as well as our field observations. Ecological and morphological traits used in this study were plant growth form, sexual system, flower size, flower color and fruit type (Table 1).

Comparison among floristic components and with other forest communities/floras

The distribution types of woody plants at the generic level were obtained from Wu (1991). After exclusion of the genera with world distribution (cosmopolitan) or endemic to China, we compared the frequency distribution of sexual systems among distribution types. Due to different reproductive traits used in other studies, it is difficult to compare all parameters except the frequencies of sexual systems of tree species. We compared the frequencies of hermaphroditic, monoecious and dioecious species of trees in SEBLF of Ailao Mountains with those of six tropical forests (Ashton 1969; Bawa et al. 1985a; Chen and Li 2008; Gross 2005; Ibarra-Manríquez and Oyama 1992; Matallana et al. 2005) and five temperate forests of North America (Bawa and Opler 1975).

Statistical analysis

Associations between sexual system and growth form, flower size, flower color and fruit type were assessed with chi-square tests. The null hypothesis for all comparisons was that the frequency distribution of sexual systems in the various traits and ecological categories was not significantly different from the distribution for the entire flora or vegetation. To take into account the phylogenetic effect on these associations, we repeated the analysis at the generic level.

RESULTS

Taxonomic representation

The survey included 703 species belonging to 103 families and 296 genera (Table S1, see supplementary material online). Ten families (Rosaceae, Lauraceae, Fagaceae, Ericaceae, Papilionaceae, Theaceae, Araliaceae, Moraceae, Myrsinaceae, Euphorbiaceae) accounted for 43.7% of all surveyed species, in which each one had >15 species. The 10 biggest genera (Rhododendron (Ericaceae), Lithocarpus (Fagaceae), Ilex (Aquifoliaceae), Castanopsis (Fagaceae), Ficus (Moraceae), Symplocos (Symplocaceae), Rubus (Rosaceae), Euonymus (Celastraceae), Eurya (Theaceae) and Acer (Aceraceae)) accounted for 20.5% of all species, whereas 171 genera had only one species.

Sexual systems

Of 103 families, 55 (53.4%) families were exclusively hermaphroditic, 13 (12.6%) were exclusively dioecious and 11(10.7%) were exclusively monoecious. The remaining 24 (23.3%) were families with at least two kinds of sexual systems (Table S1, see supplementary material online). Of 296 genera, 190 (64.2%) were hermaphroditic, 60 (20.3%) were dioecious and 43 (14.5%) were monoecious. A genus was often exclusively of one type of sexual system except in three genera: Ficus (Moraceae), Symplocos (Symplocaceae) and Piper (Piperaceae). Of 703 species, 423 (60.2%) were hermaphroditic, 169 (24.0%) were dioecious and 111(15.8%) were monoecious.

Growth forms and sexual systems

There were 279 trees (39.7%), 362 shrubs (51.5%) and 62 climbers (8.8%). The distribution of sexual systems differed significantly among plant growth forms both at the species and generic levels (χ² = 76.95, P < 0.001 and χ² = 10.30, P < 0.05 respectively; Table 2). Hermaphroditism was more represented among shrubs and less represented among trees. Monoecy was more represented among trees and less represented among shrubs and climbers. The relative frequency of dioecy among growth forms did not differ significantly.

Flower sizes and sexual systems

Among 703 species, 381 (54.2%) species produced small flowers, 206 (29.3%) with large flowers and 116 (16.5%) with medium flowers. Sexual systems were significantly associated with flower sizes both at the species and generic levels (χ² = 203.96, P < 0.001 and χ² = 89.57, P < 0.001, respectively; Table 2). Monoecious and dioecious species produced proportionately more small flowers and less large and medium flowers, while hermaphroditic species produced proportionately more large flowers and less small and medium flowers.

Flower colors and sexual systems

The information of flower colors was only available for 641 of the 703 species. Among them, 257 (40.1%) species produced white flowers, 136 (21.2%) species with green flowers, 121 (18.9%) with yellow species, 79 (12.3%) species with red flowers and 48 (7.5%) species with purple flowers. Sexual systems were significantly associated with flower colors both at the species and generic levels (χ² = 119.59, P < 0.001 and χ² = 64.72, P < 0.001, respectively; Table 2). Dioecious and monoecious species produced proportionately more green and yellow flowers, while hermaphroditic species produced proportionately more white, red and purple flowers.

Fruit types and sexual systems

A total of 414 (58.9%) species produced fleshy fruits and 289 (41.1%) species produced dry fruits. Sexual systems and fruit types were significantly associated both at the species and generic levels (χ² = 86.85, P < 0.001 and χ² = 16.23, P < 0.001 respectively; Table 2). Dioecious species produced proportionately more fleshy fruits and less dry fruits, while monoecious species produced proportionately more dry fruits and less fleshy fruits. Hermaphroditic species produced slightly more dry fruits and less fleshy fruits.

Comparison among floristic components and with other communities/floras

Among 290 genera, which we got the information of their distribution types, 186 (64.1%) genera were tropical distribution and 90 (31.0%) genera were temperate distribution. Sexual systems were significantly associated with distribution types at the generic level (χ² = 9.90, P < 0.05). Monoecy was more represented among temperate genera and less represented among tropical genera (χ² = 8.20, P < 0.05). Dioecy and hermaphroditism were slightly more represented among tropical genera and less represented among temperate genera (χ² = 0.98, P > 0.05 and χ² = 0.68, P > 0.05, respectively).

The trees of SEBLF at Ailao Mountains had a lower proportion of hermaphroditic species and higher proportion of monoecious species compared with tropical forests of SW China, Malaysia, Mexico, Costa Rica, Brazil and Australia. The proportion of dioecious species of trees at Ailao Mountains was slightly lower than those of tropical forests in SW China, Malaysia, Mexico and Brazil, but higher than that of tropical northern Australia. The proportion of hermaphroditic and dioecious species of trees of SEBLF in Ailao Mountains was higher than those of five temperate forests in North America (Table 3), while temperate forests in North America had a much higher proportion of monoecious species.

DISCUSSION

SEBLF at Ailao Mountains had a high proportion of diclinous sexual system in which 24.0% of the species were dioecious and 15.8% of the species were monoecious. Dioecy was significantly associated with fleshy fruits and monoecy was significantly associated with dry fruits in the forest. Monoecy was more represented in temperate components and less represented in tropical components. Species with small and inconspicuous flowers were common in SEBLF at Ailao Mountains.

High incidence of diclinous sexual systems in the SEBLF

The percentage of dioecious species (24.0%) in the forests of Ailao Mountains was much higher than the estimation for the entire angiosperm flora (6%) (Renner and Ricklefs 1995), but was comparable with that of in tropical forests (Bawa et al. 1985a; Bullock 1985; Chazdon et al. 2003; Ibarra-Maríquez and Oyama 1992). Previous studies have shown that dioecy was more commonly represented in tropical than temperate forest environments (Bawa 1974, 1980; Bawa and Opler 1975; Flores and Schemske 1984; Renner and Ricklefs 1995). High percentage of dioecious species of SEBLF in Ailao Mountains may reflect the tropical origin of the flora (Wu 1965). After the uplift of Hengduan Mountains (including Ailao Mountains) along with Tibet Plateau, the nearly north-south direction ridges and valleys may facilitate the species dispersal between warm temperate region in the north and tropical region in the south (Fig. 1). The forests were evolved to including some temperate elements, but the main components are still tropical at the generic level. Considerable temperate components and their association with monoecious sexual system may account for the high percentage of monoecious tree species in SEBLF, which is higher than many tropical forests (Bawa et al. 1985a; Bullock 1985; Chazdon et al. 2003; Gross 2005; Ibarra-Maríquez and Oyama 1992).

Fruit types and the evolution of plant sexual systems

In woody species of SEBLF at Ailao Mountains, fleshy fruits were common and were significantly correlated with dioecy. This association is common for many tropical forest communities (Carpenter et al. 2003; Geber et al. 1999; Ibarra-Manríquez and Oyama 1992), but is rarely documented in SEBLF. One explanation for this correlation was that dioecy might be more easily established in animal-dispersed species because female fitness could be enhanced disproportionately by increasing reproductive effort (Bawa 1980, 1982; Givnish 1980). Meanwhile, dry fruits were significantly associated with monoecy in woody species of SEBLF at Ailao Mountains, an association rarely documented in other forest communities (but see Flores and Schemske 1984; Gross 2005). However, in gymnosperms, almost all monoecious species produced wind-dispersed cones and almost all dioecious species produced animal-dispersed fleshy fruits (Givnish 1980). In wind-pollinated gymnosperms, the male effort was relatively high (vast amount of pollen) and most additional pollen fell near the parents and competed for fertilizing the ovules of the same individual, but an increasing investment in ovules and fruits would produce a disproportionately large increase in seeds timely dispersed by animals. In such a case, females could invade a monoecious population and sexually dimorphic populations were established (Givnish 1980). Although most woody angiosperm species are not wind-pollinated (personal observation), there is still a significant correlation between monoecious sexual system and dry fruit type in SEBLF of Ailao Mountains. We hypothesize this correlation might be due to the trade-off of resource allocation between male and female functions. Fruits of many monoecious species, such as Fagaceae, Juglandaceae and Betulaceae, could not mature until next summer or autumn in the forests of Ailao Mountains (Qiu 1998). This might be related to high altitude and relatively low temperature in the mountain forests. However, maintaining the fruits for a long period would be a severe resource burden for mother plants. In such situation, it was very costly for monoecious species to produce and maintain nutrient rich fleshy fruits. Monoecious plants might increase female fitness through producing less costly dry fruits.

The association between monoecious sexual system and dry fruit type may also be generated spuriously by their phylogenetic structure. For example, the single family Fagaceae accounted for more than one-third of the monoecious species and all produce dry fruits at Ailao Mountains. Several studies have shown that reproductive traits are highly phylogenetically structured in tropical forests (Chazdon et al. 2003; Chen and Li 2008; Gross 2005). Although we made phylogenetic correction at the generic level, further studies should investigate the phylogenetic effect at higher taxonomic level.

Floral traits and presumed pollination systems

Small and inconspicuous flowers dominated the flora of Ailao Mountains, which suggested that generalized pollination systems may be common. However, pollination biology at the community level has not been investigated in any detail at Ailao Mountains. In the Oriental Region, relatively specialized pollinator groups such as large solitary bees, moths and vertebrates were relatively less important, while social bees, beetles, other bees and flies were more important pollinator groups in comparison with Neotropics (Corlett 2004). Generalized pollination was proposed to promote the evolution of sexual dimorphism (Bawa 1980, 1994; Bawa and Opler 1975; Lloyd 1982) and might account for the high percentage of diclinous sexual systems in several communities in the region (Ashton 1969; Chen and Li 2008; and therein). However, in a middle elevation wet evergreen forest in Western Ghats, India, ∼75% of the species were specialized to a single pollinator group such as bees, beetles or moths (Devy and Davidar 2003), and therefore more specialized pollination systems were common in that region. The degree of generalization versus specialization in pollination systems and its significance for the evolution of sexual systems cannot be identified unless community-level surveys are conducted. Current work on species in SEBLF of Ailao Mountains is aimed at identifying the extent to which pollination biology plays a significant role in the evolution of plant sexual systems.

SUPPLEMENTARY DATA

The number of genera, species and percentage of species with hermaphroditic, dioecious, monoecious systems and dry, fleshy fruits in each family are available at Journal of Plant Ecology online.

FUNDING

National Natural Science Foundation of China grant 30225007 and the Fund for Top One Hundred Young Scientists to Q.-J.L.

We thank the assistants in Ailaoshan Forest Ecosystem Research Station (affiliated station of China Ecological Research Network) for their kind reception and help in the field; Mr P.-Y. Ren for preparing the figure; thanks to Professor H. Zhu and Dr J.-P. Shi for providing their plot data and valuable suggestions; curators of HITBC and KUN for permitting and helping us to examine specimens and S. Martén-Rodríguez for comments on the earlier versions of the manuscript.

References