southern Yunnan, China - Universität Hohenheim

Transcription

southern Yunnan, China - Universität Hohenheim
UNIVERSITÄT HOHENHEIM FACULTY OF AGRICULTURAL SCIENCES INSTITUTE OF PLANT PRODUCTION AND AGROECOLOGY IN THE TROPIC AND SUBTROPICS Agroecology in the Tropic and Subtropics
apl Prof. Dr. Konrad Martin Animal‐plant‐interactions at different scales in changing tropical landscapes of southern Yunnan China Dissertation Submitted in fulfilment of the requirements for the degree “Doktor der Agrarwissenschaften” (Dr. Sc. Agr.) to the Faculty of Agricultural Sciences Presented by Ling‐Zeng Meng Stuttgart, Germany 2011 This thesis was accepted as a doctoral dissertation in fulfilment of the requirements for the degree“Doktor der Agrarwissenschaften” by the Faculty of Agricultural Sciences at Universität Hohenheim on January 27, 2012 Date of oral examination: February 13, 2012 Examination Committee Supervisor and Reviewer: apl Prof. Dr. Konrad Martin Co‐Reviewer: Prof. Dr. Claus P.W. Zebitz Additional examiner: Prof. Dr. Martin Dieterich Dean and Head of Committee: Prof. Dr. Dr. Rainer Mosenthin CONTENT CHAPTER 1: General Introduction………………………………………………………………………………1 Problem statement: Threats to biodiversity in tropical forest regions and its appraisal………………………………………………………………………………………………2 The study region……………………………………………………………………………………….5 Study outline and specific objectives………………………………………………………14 References………………………………………………………………………………………………18 CHAPTER 2: Carabid beetle communities and species distribution in a changing tropical landscape (southern Yunnan, China)………………………………………………………..23 CHAPTER 3: Contrasting responses of hoverflies and wild bees to habitat structure and land use change in a tropical landscape (southern Yunnan, SW China)…………..43 CHAPTER 4: Spatial and temporal effects on seed dispersal and seed predation of Musa acuminata in southern Yunnan, China…………………………………………………………65 CHAPTER 5: Young leaf protection in the shrub Leea glabra in south–west China: the role of estrafloral nectaries and ants…………………………………………………………………….85 CHAPTER 6: General Discussion……………………………………………………………………………..101
SUMMARY……………………………………………………………………………………………………………..109 ZUSAMMENFASSUNG ………………………………………………………………………………………………..113 ACKNOWLEDGEMENT...………………………………………………………………………………………………117 PUBLICATION LIST…………………………………………………………………………………………………....119 CURRICULUM VITAE…………………………………………………………………………………………………. 120 Chapter 1 General Introduction Chapter 1 General Introduction 1
Chapter 1 General Introduction 1 General Introduction 1.1 Problem statement: Threats to biodiversity in tropical forest regions and its appraisal Tropical forests harbor the highest species diversity among the terrestrial ecosystems. Large areas of diverse tropical forest are worldwide lost or degraded every year with dramatic consequences for biodiversity. The main causes of tropical forest biodiversity loss are human‐induced processes, including deforestation and fragmentation, over‐exploitation, invasive species and climate change (Morris 2010). Approximately half of the earth's closed‐canopy tropical forest has already been converted to other uses, and the population of tropical countries, having almost trebled since 1950, is projected to grow by a further 2 billion by 2030 (Wright 2005). The Southeast Asia is experiencing the highest relative rates of deforestation and forest degradation in the humid tropics due to agricultural expansion, logging, habitat fragmentation and urbanization, which are expected to result in species declines and extinctions. In particular, growing global demands for food, biofuel and other commodities are driving the rapid expansion of large monocultures of oil palm and rubber plantations at the expense of lowland dipterocarp forests, further jeopardizing Southeast Asian forest biotas. Furthermore, Southeast Asia contains the highest mean proportion of country‐endemic bird (9%) and mammal species (11%) and also has the highest proportion of threatened vascular plant and reptile species among the tropical forest regions (Sodhi et al. 2010). Given the rate of destruction and the high concentration of endemic species in the region, Southeast Asia could lose 13‐42% of local populations by the turn of the next century, at least 50% of which could represent global species extinction (Koh & Sodhi 2010). To estimate consequences of forest degradation and fragmentation for natural diversity and species composition in tropical forest areas, it is not enough to study the inventories of the original forest alone. As pointed out by Lewis (2009), habitat ‘loss’ in tropical forests is only one part of the picture. The outcome of tropical deforestation is rarely the complete destruction of a given area by leaving ‘nothing’, 2 Chapter 1 General Introduction but rather an altered, partly degraded and fragmented landscape. Depending on the causes and intensity of deforestation, the activities are leaving landscapes of habitat patches, which may include agricultural land, plantations, and various stages of natural succession leading to secondary forests. Such habitats may represent a matrix separating relatively intact plots of undisturbed old‐growth forest, but the current knowledge of the value of such matrix habitats for biodiversity conservation is still limited. Existing studies showed a wide range of different results, suggesting that the responses of tropical forest species to habitat change are highly idiosyncratic (Lewis 2009). For example, Barlow et al. (2007) examined the conservation value of tropical primary, secondary, and plantation forests for 15 taxonomic groups and found that different taxa varied markedly in their response to patterns of land use in terms of species richness and the percentage of species restricted to primary forest (varying from 5% to 57%). Although they showed that areas of native regeneration and exotic tree plantations can provide complementary conservation services, Barlow et al. (2007) also provide clear empirical evidence demonstrating the irreplaceable value of primary forests. Perfecto and Vandermeer (2008) argue that agricultural landscapes should be an essential component of any conservation strategy in tropical forest regions. They present evidence that many tropical agricultural systems have high levels of biodiversity and that these systems represent not only habitats but also a high‐quality matrix that permits the movement of forest organisms among patches of natural vegetation. Agro‐forestry systems were also considered to play an important role in biodiversity conservation in human‐dominated tropical landscapes (Bhagwat et al. 2008). However, Lewis (2009) argues that landscape fragmentation leads to increasing ‘biotic homogenization’: Although disturbed forests may often have an equal or even a greater number of species than undisturbed forests, these species are typically drawn from a restricted pool; and endemic, restricted‐range or habitat‐specialist species are most likely to decline or go extinct. There is also number of studies that confirm significant reductions of fauna in uniform plantations such as oil palm and rubber compared to natural forest. For example, Danielsen and Heegaard (1995) found that conversion of primary forest to rubber and oil palm in Sumatra led to simple, species‐poor and less diverse animal communities with fewer specialized species and fewer species of importance to conservation. In the plantations, only 5‐10% of the primary‐forest bird species were recorded. Primates, squirrels and tree‐shrews disappeared except for one species. Similarly, Peh et al. 3
Chapter 1 General Introduction (2005) found reductions in primary‐forest species of more than 70% in such habitat types in Malaysia. Most studies investigating these threats have focused on changes in species richness or species diversity. However, to understand the absolute and long‐term effects of anthropogenic impacts on tropical forests, research should also consider functional diversity. That is, the interactions between species and guilds, their organization, and structural attributes of communities (Lewis 2009; Morris 2010). All species are embedded in complex webs of mutualistic and antagonistic interactions, which are most complex and diverse in tropical forests (Reagan & Waide 1996). Therefore, species extinctions can be a direct result of environmental changes such as habitat loss, and the interaction (link) between species can be lost when one of the partners disappears (Tylianakis et al. 2010). Therefore, understand the implications of biodiversity loss, it is crucial to analyze relationships between species richness and ecosystem functions at different scales. Ecosystem functions refer to the processes of basically natural ecosystems which are essential for the maintenance of the structure of the ecosystem type (e.g., forest), the biodiversity and the biotic and abiotic cycles of the respective system. They include feeding relationships between species (food webs), primary and secondary production and the flows of matter and energy, and the level of function in the landscape context (e.g. Díaz et al. 2007; Clavel et al. 2011). Finally, losses or changes in functional diversity may also affect important ecosystem services, i.e. benefits that people obtain from ecosystems (Costanza et al. 1997), such as crop pollination, the use of wild plant species for food and medicine or fiber production. Generally, the importance of functional interactions in relation to biodiversity has been detected in various studies. High plant species diversity depends on a highly diverse pollinator community and vice versa (e.g. Fontaine et al. 2006), and rain‐forest fragmentation can reduce the abundance and diversity of frugivores, which in turn is likely to reduce the dispersal of a certain plant taxa and may alter patterns of plant regeneration in forest fragments (Moran et al. 2009). Other important mutualistic interactions in tropical forest ecosystems are given by the various types of ant‐plant‐relationships. In addition, there a many studies showing that and how spatial configuration, habitat structure, patch size and the degree of isolation of habitats affect the functional diversity in fragmented tropical landscapes (e.g. Tscharntke et al. 2008). 4 Chapter 1 General Introduction The overall objective of the study entitled “Animal‐plant‐interactions at different scales in changing tropical landscapes of southern Yunnan, China” was to contribute to the knowledge on species interactions and functional diversity in a fragmented tropical landscape of southern Yunnan, China, at different scales. Studies included (a) the species richness of ground beetle (Carabidae) communities in relation to land use type and habitat structure within a landscape mosaic, (b) primary and secondary seed dispersal and seed predation in the wild banana species Musa acuminata at different sites, (c) diversity of insect pollinators (wild bees and hoverflies) in relation to habitat type and flowering resources and (d) mutualistic interactions between the forest understorey shrub Leea glabra and ants. 1.2 The study region The study area was located in the Autonomous Prefecture of Xishuangbanna in the southern Yunnan province, southwest China (Fig. 1.1). Fig. 1.1: The geographical location of Xishuangbanna and the research sites. (A) Plot of tropical seasonal forest near the Xishuangbanna Tropical Botanical Garden (XTBG); (B) Plot of tropical seasonal lowland rain forest at Mengla Natural Reserval (MNR); (C) Plot of tropical seasonal rainforest at Naban River Watershed National Nature Reserve (NRWNNR). 5
Chapter 1 General Introduction 1.2.1 Geography and climate The tropical area of southern China is climatically and biogeographically located at the northern edge of tropical Asia, including southeastern Xizang (Tibet, lower valleys of the southern Himalayas), southern Yunnan, southwestern Guangxi, southern Taiwan and Hainan Island. Southern Yunnan is the largest tropical area of China still covered by forests (Zhu et al. 2006). Xishuangbanna is located between 21°08′ ‐ 22°36′N and 99°56′ ‐ 101°50′E and covers an area of 19.200 km2. It borders Myanmar in the southwest and Laos in the southeast. The region has a mountainous topography, with ridges running in a north–south direction, decreasing in elevation southward. The elevation ranges from 491 m in the lowest part of the Mekong River valley in the south to mountains of ca. 2400 m in the north. The uplift of the Himalayas leads to the penetration of warm and moist tropical air masses from the Indian Ocean to Xishuangbanna in summer, and forms a barrier preventing cold air masses from the north reaching the region in the winter, i.e. a typical monsoon climate (Zhang 1986). These conditions allow the existence of a tropical rain forest in its altitudinal and latitudinal northern limits. The Mekong River runs through the region from northwest to southeast (Cao et al. 2006). The monsoon climate is characterized by two main seasons over the year. Between May and October, the tropical Southwest Monsoon from the Indian Ocean accounts for about 80% of the annual rainfall, whereas the dry and cool air of the southern edges of the subtropical jet streams dominates the climate between November and April. The dry season from November and April is further divided into a cool–dry (November to February) and a hot–dry (March to April) period. Annual precipitation is higher in the eastern part of the region than in the west with a total annual average between 1500‐1600 mm. The mean annual temperature is around 22‐23° C. It is highest in May/June and lowest in December/January (Cao et al 2006). A typical climate diagram from Xishuangbanna is shown in (Fig. 1.2). There are three main soil types in the region. Laterite soils developed from siliceous rocks, such as granite and gneiss and occur between 600–1,000 m elevation with a deep solum but a thin humus horizon. The lateritic red soil from the rock substrate of sandstone occurs in the area above 1,000 m elevation. Limestone hills have soil from rock substrate of hard limestone of Permian origin with a pH of 6.75. The tropical rain 6 Chapter 1 General Introduction forest of Xishuangbanna occurs mainly on laterite soils with pH values of 4.5–5.5. A small portion of the tropical rain forest occurs on limestone soils (Zhu et al. 2006). Fig. 1.2: Distribution of monthly precipitation and temperature of study area (average of 40 years, 1959‐1998) ©by Xishuangbanna Forest Ecology Station 2002). 1.2.2 Natural vegetation and biodiversity The present tropical rain forest in Xishuangbanna is at the altitudinal and latitudinal limits of tropical rain forests in the northern hemisphere. The tropical rain forest in Xishuangbanna was classified into two subtypes, i.e. a tropical seasonal rain forest in the lowlands and a tropical montane rain forest at elevations from 900‐1600 m. The tropical seasonal rain forest occurs up to an elevation of 900 m and has 3–4 indistinct tree layers, with tree heights of 30‐60 m in the top layer and about 30 percent crown coverage. The second layer is the main canopy layer, and is up to 30 m tall with an almost continuous canopy (70–80% coverage), and greatest stem density. The third layer is 5–18 m tall with a crown cover of about 40%, consisting of small trees and juveniles of species from the upper layers. Buttresses and cauliflory are common, and 7
Chapter 1 General Introduction both big woody lianas and vascular epiphytes are abundant. The forest is mainly evergreen despite the fact that there are some deciduous trees in the canopy layer. Further subtypes are the tropical seasonal moist forest on the middle and upper limestone slopes with shallow soils at elevations of 650‐1300m with two distinct tree layers, and the tropical montane evergreen broad‐leaved forest on mountain slopes and summits above 1000m and valleys above 1300m with two conspicuous tree layers, of which the top layer is 15–25m tall with a dense crown coverage. The tropical montane evergreen broad‐leaved forest occurs at elevations from 900‐1600m. Another distinct forest type is the tropical deciduous Monsoon forest with a mosaic distribution within seasonal rain forest. It represents a transitional forest type between seasonal rain forest and savanna. In Xishuangbanna, Monsoon forests occur on the banks of the Mekong River and in wide basins characterized by seasonal drought. The Monsoon forest is usually 20–25m tall with 1–2 deciduous tree layers. Woody lianas and epiphytes are scarce. (Zhu et al. 2006; Zhu 2008). In total, the primary vegetation of Xishuangbanna can be classified into four main vegetation types, i.e. tropical rain forests, tropical seasonal moist forests, tropical montane evergreen broad‐leaved forests and tropical monsoon forests, including two vegetation sub‐types, four formation groups and eighteen formations, which were further described by Zhu et al. (2006), Zhang and Cao (1995)and Lü et al. (2010). The floristic similarities between the flora of southern Yunnan and those of tropical Asia are more than 80% at the family level and more than 64% at the generic level. This suggests that the tropical flora of southern Yunnan has a close affinity with tropical Asian flora and supports the idea that the flora of southern Yunnan, together with mainland Southeast Asian flora, belongs to the Indo‐Malaysian floristic subkingdom of the Paleotropical kingdom (Zhu 2008). Yunnan including tropical Xishuangbanna and its adjacent tropical forest regions represent the most diverse eco‐region of China, which is part of the “Indo‐Burma hotspot”, one of the 34 global hotspots exceptionally rich in biodiversity (Biodiversity Hotspots 2007). Although Xishuangbanna represents only 0.2% of the area of China, it contains more than 5000 species of higher plants (16% of the nation’s total), 102 species of mammals (21.7%), 427 species of birds (36.2%), 98 species of amphibians and reptiles (14.6%), and 100 species of freshwater fish (2.6%) (Zhang & Cao 1995). 8 Chapter 1 General Introduction 1.2.3 Land use and land use change Xishuangbanna is also known for its cultural diversity, represented by more than 10 ethnic groups with different traditions and land use. Among the largest groups, the Dai live in lowland areas near rivers, with rice cultivation as their major agricultural activity. The Hani live in mountainous areas practicing slash‐and‐burn farming, whereas the Lahu are or were mainly hunters. Previous to the land use changes in the recent decades, the local economy used to depend largely on traditional agriculture, including rice production in the lowlands and slash‐and‐burn farming on mountainous slopes. Land use was characterized by interactions of mixed systems, in which forests were maintained by upland communities to supply clean water, timber, and non‐timber forest products, as well as hunting and fishing. In the beginnings in the 1950s, the first plantations of rubber (Hevea brasiliensis) became established in Xishuangbanna. At that time, rubber production was integrated into traditional land use systems and cultivated in mixed‐cropping systems, which largely were substituted by monocultures after rubber turned into a high benefit cash crop within the last 20‐30 years (Wu et al. 2001). Between 1999 and 2007, the world production of natural rubber increased by 35%, from 6.7 to 10.3 million tons per year. (FAOSTAT 2009). Within the last decade, continued expansion of rubber cultivation in Xishuangbanna has mainly been driven by an expanding rubber market, powered largely by the modernization and massive increase of vehicles in China. In addition, free market and the lure of cash products have encouraged numerous private landholders to consolidate and turn to rubber over the last two decades. Rubber production accounted for more than 30% of the regional economy in 2007, with an expected long‐term increase (Liu et al. 2006). Consequently, land use and vegetation cover of Xishuangbanna changed dramatically within the last decades. From the analysis of satellite images, Li et al. (2007) found that in 1976, forests covered approximately 70% of Xishuangbanna, but in 2003, they covered less than 50%. Tropical seasonal rain forest was the forest type most affected by the expansion of rubber plantations, and a total of 139.500 ha were lost. In 1976, forest was the dominant land cover category at all elevations, but by 2003, rubber plantations dominated areas below 800 m, representing the most suitable elevation for rubber cultivation. In total, most of the land below 800‐1000 m asl is already covered by rubber plantations. Furthermore, the number of forest fragments 9
Chapter 1 General Introduction increased from 6,096 to 8,324, and the mean patch size declined from 217 to 115 ha. In addition, an increase in arable land and especially shrublands between 800 and 1300m also reduced forest cover. Hu et al. (2008) observed the land use change in the Menglun area, a representative region of Xishuangbanna, from 1988‐2006. The results showed that over the 18‐year period, rubber plantations increased from 12% of total land cover to 46%, while forested area decreased from 49 to 27%, and swidden fields from 13 to 0.5%. A typical situation of the present land use of Xishuangbanna at lower elevations is shown from the Naban River valley, where most of the present studies have been conducted. The valley (ca. 11,000 ha) is a tributary river of the Mekong and located within the area of the Naban River Watershed National Nature Reserve (NRWNNR) in Xishuangbanna. As shown in Fig. 1.3 and 1.4, most of the valley area is covered by rubber plantations at elevations between 550‐1000m. However, rubber plantations do not represent a uniform type of land use, but rather a spatio‐temporal dynamic system, ranging from young and open to closed canopy stands of very different ecological conditions and plant species. Stands of different age exist at the same time within a rotati on cycle of about 40 years, when the latex production of the trees decreases. Then, the plantations are clearfelled and become substituted by new rubber saplings. The remaining land use types in the valley include secondary and primary forest fragments, grassland and shrubland successions as well as rice fields in the valley bottom along the river. Other field crops are produced around the small villages. A characteristic section of the landscape is shown in Fig. 1.5. Different traps used in this study are shown in Fig. 1.6. 10 Chapter 1 General Introduction Fig. 1.3: Land use map of the Naban River Watershed National Nature Reserve (NRWNNR). The low elevations of the Naban River valley are dominated by rubber plantations. Figure1.4: Major land use systems presently found in the Naban River valley. 11
Chapter 1 General Introduction Fig. 1.5: The landscape of different land‐use type at Naban River Watershed National Natural Reserve (NRWNNR). (a) a typical landscape within NRWNNR, (b) old rubber plantation at Anmaxinzhai, (c) young rubber plantation at Mandian, (d) rice field at Guomenshan, (e) young rubber plantation at Naban, (f) up valley of Naban River mixed with rice field, (g) forest edge at Naban, (h) forest landscape at Guomenshan (Photographs by Jing‐Xin Liu). 12 Chapter 1 General Introduction Fig. 1.6: Different insect trap types used in this study. (a, b) Malaise trap in the forest understorey, (c) aerial eclector in young rubber plantation, (d) aerial eclector in the forest canopy, (e) pitfall trap in the forest understorey, (f) Malaise and pitfall traps in open land. 13
Chapter 1 General Introduction 1.3 Study outline and specific objectives Despite the enormous contribution of invertebrates to global biodiversity and ecosystem function, knowledge on the patterns and causes of insect responses to tropical rainforest destruction and fragmentation is limited compared to that for other taxa (McGarigal & Cushman 2002; Grimbacher et al. 2006). To evaluate the effects of forest fragmentation on biodiversity, is essential not to consider species inventories of the remaining forest patches alone, but also the managed ecosystems of the landscape matrix to understand how they influence components of biodiversity. (Perfecto & Vandermeer 2008). Landscape‐level analyses have led to the suggestion that fragment size by itself may have little to do with the species present in the fragments; rather, patterns at the patch scale may be driven by the amount of habitat lost at the landscape scale (Fahrig 2003). Generally, human‐modified matrix habitats may have affect species diversity at the landscape scale in three ways. First, they may facilitate movement among patches by acting as a corridor between habitats, second they may act as barriers leading to increased isolation or third, they may provide alternative or secondary habitats for species outside the original forest depending on the habitat quality. Studies on beetle diversity in fragmented tropical forest landscapes showed that a variety of factors may influence species distribution patterns. Referring to soil‐dwelling beetles, Didham et al. (1998) found that species composition changed significantly and independently with both decreasing distance from forest edge and decreasing fragment size in the study area of Amazon. Beetle diversity was affected by different environmental factors including temperature, vegetation structure, litter biomass and moisture. They also concluded that common but poorly dispersing species are the first to become extinct due to habitat destruction. Rare species with a higher dispersal ability have higher changes to persist. Other studies on soil‐dwelling beetles showed a decreasing species richness with increasing disturbance (Goehring et al. 2002; Grimbacher et al. 2006) In this context, the study of Chapter 1 was conducted to investigate the ground beetle (Carabidae) communities of the major land use types in the Naban River valley in order to identify species richness and its relation to land use type and management, 14 Chapter 1 General Introduction including forest plots, rubber plantations of different age, grass‐ and shrub‐land as well as rice‐field areas. Another group of insects considered in many studies of fragmented tropical landscapes are pollinators, such as wild bees and hoverflies. Most research was conducted under the aspect of the provision of pollination as an important ecosystem service for both natural plants and cultivated crops. For tropical crops, a list of 1330 species including their potential pollinating taxa, indicates that about 70% of tropical crops benefit from animal pollination (Roubik 1995), and many tropical crops (e.g., cocoa, passion fruit, dragon fruit, vanilla, oil palm) depend on pollination by naturally occurring insects such as wild bees and wasps. However, pollination services are considered to be at a risk (e.g., Gallai et al. 2009). The two major threats for pollinator diversity and crop pollination services include the destruction and fragmentation of natural and semi‐natural habitats, and the intensification of agricultural landscapes (Steffan‐Dewenter & Westphal 2008). Moreover, the distance of insect‐pollinated plant populations to high‐quality habitats for pollinators affects their pollination and reproductive success. However, the effect of landscape composition on large‐scale foraging routes of pollinators, and hence landscape‐wide pollen dispersal and cross‐pollination rates are largely unknown. A study dealing with effects of tropical forest fragmentation on pollinator communities was conducted by Brosi et al (2008). In Costa Rica, they examined bee community responses to forest fragment size, shape, isolation and landscape context by sampling foraging bees. They found strong changes in bee community composition, which was also strikingly different between forests and pastures, despite their spatial proximity. Their results agree broadly with other studies that have found contrasting responses to habitat fragmentation from different bee groups. Overall, crop visitation rates decline with increasing distance from pollinator habitats, as demonstrated in a meta‐analysis of 16 case studies in tropical and temperate regions (Ricketts et al. 2008), illustrating the potential threat for pollinator ecosystem services. They found strong exponential declines in both pollinator richness and native visitation rate. Visitation rate drops more steeply in tropical compared with temperate regions. In this context, the study of Chapter 2 was conducted to investigate species diversity and distribution patterns of hoverflies and wild bees in relation to land use change in the Naban River valley. The main objectives were (a) to identify habitat types of 15
Chapter 1 General Introduction major importance for the existence and preservation of the two pollinator taxa within the landscape, considering natural forest plots, rubber plantations of different age, grassland and agricultural land, and (b) to find out if and how richness of flowering herb species in these habitats is related to species diversity and abundance of hoverflies and wild bees. For wild plants, not only pollination, but also seed dispersal is a relevant factor for survival and distribution of the populations. Similar to animals, the chance for tropical forest species to persist outside the original forest strongly depends on the quality and management of the matrix habitats. However, about 80% of the woody plant species of tropical forests depend on animals as seed dispersers (Jordano 1992), which also may be affected by forest fragmentation. Furthermore, seed fate and seedling recruitment is also affected by animal seed predators, which may be more or less abundant in the matrix compared to the forest habitat. Overall, the process of seed dispersal is complex and covers basically two categories or phases (Vander Wall et al. 2005a). The first category or phase (primary dispersal) includes the seed removal from the parent plant, usually by vertebrate frugivores, which then release the seeds by defecation and regurgitation. Frugivorous animals clearly facilitate tropical forest regeneration and help to maintain species diversity by introducing seeds from outside disturbed areas and help to promote the regeneration of late successional plant species. Overall, the population and community dynamics of tropical forests would likely be very different in the absence of frugivorous bats (Muscarella & Fleming 2007). However, results from a study by López‐Bao and González‐Varo (2011) also suggest that a strong preference for cultivated fruits by mammal frugivores may influence their spatial foraging behavior and lower their dispersal services to wild species. The second category or phase (secondary dispersal) includes the processes leading to the end‐point of the seed and may include dispersal mechanisms by animals on the ground (insects, especially ants, rodents and birds), which remove seeds from feces or from the soil (Herrera 2002; Vander Wall et al. 2005b). Together with secondary dispersal, seed predation is another important process which affects the final seed fate (Wang & Smith 2002). In this context, the study of Chapter 3 was conducted to contribute to further understanding of animal‐seed interactions by the analysis of processes affecting the fate of seeds of a wild banana species (Musa acuminata) in forest and shrubland 16 Chapter 1 General Introduction habitats under consideration of different temporal (seasons) and spatial (site and habitat) factors. Besides pollination and seed dispersal, many tropical plant benefit from animals, specifically ants through the protection from herbivores. Numerous examples showed that ants can act as biotic defenses, protecting plants against herbivorous animals, mainly insects. In return, plants offer benefits such as shelter (nesting sites termed domatia) and food rewards (extrafloral nectar or food bodies. Two strategies can be distinguished within defensive ant–plant interactions. Ant–plants, or myrmecophytes, are continuously inhabited by specific ant species most of their lifetime, whereas non‐domatia‐bearing plants do not house interacting ants (myrmecophiles) and gain protection from a facultative and opportunistic ant community (Chamberlain & Holland 2009; Rosumek et al. 2009). In facultative associations, plants secrete small volumes of nectar from organs in various locations outside the flowers, termed extrafloral nectaries (EFNs). The EFNs are highly diverse in structure and ontogeny, and are found on an array of vegetative and reproductive parts, most commonly on developing structures including new leaves, flowers and fruits. They are easily accessible to plant visitors, and a variety of insects, including diverse ant species, feed there on a regular basis. By exhibiting aggressive behavior towards herbivores, EFN‐gathering ants can positively affect plant fitness by decreasing herbivore damage to leaves buds or flowers. The production of ant attractants has been predicted to be more common in plants with short‐lived leaves, because a continuous investment is needed to feed the ants, and as a leaf ages this cost eventually exceeds that of investing in quantitative defenses (e.g. Cogni et al. 2003; Heil & McKey 2003). Although a positive relation between the number of ants attracted and the effectiveness of defense is intuitively evident, few studies, besides the one using inducibility of EFN secretion, empirically demonstrate this relation. Even less is known about how differences in plant rewards influence species identity of attracted ants and how this in turn affects protective benefit. More studies are therefore required to determine whether the protective effect of EFN secretion varies with the quantity and/or quality of EFN (Heil & McKey 2003). In this context, the study of Chapter 4 with the forest understorey shrub Leea glabra and the defensive function of the EFNs established at the petioles of the leaves in relation to their interactions with ants under different conditions. Field experiments with naturally grown L. glabra plants were conducted to find out whether there are 17
Chapter 1 General Introduction differences in the quantity of EFNs secretion and sugar concentration between artificially damaged young and old leaves and the effects on ant recruitment and abundance on the plants following the different damage treatments. 1.4 References Barlow J, Gardner TA, Araujo IS et al. (2007). Quantifying the biodiversity value of tropical primary, secondary, and plantation forests. Proceedings of the National Academy of Sciences of the United States of America, 104, 18555‐18560. Bhagwat SA, Willis KJ, Birks HJB, Whittaker RJ (2008). Agroforestry: a refuge for tropical biodiversity? Trends in Ecology & Evolution, 23, 261‐267. Biodiversity Hotspots (2007). Available from URL http://www.biodiversityhotsp ots.org/ Brosi BJ, Daily GC, Shih TM, Oviedo F, Duran G (2008). The effects of forest fragmentation on bee communities in tropical countryside. Journal of Applied Ecology, 45, 773‐783. Cao M, Zou XM, Warren M, Zhu H (2006). Tropical forests of Xishuangbanna, China. Biotropica, 38, 306‐309. Chamberlain SA, Holland JN (2009). Quantitative synthesis of context dependency in ant‐plant protection mutualisms. Ecology, 90, 2384‐2392. Clavel J, Julliard R, Devictor V (2011). Worldwide decline of specialist species: Toward a global functional homogenization? Frontiers in Ecology and the environment, 9, 222‐228. Cogni R, Freitas AVL, Oliveira PS (2003). Interhabitat differences in ant activity on plant foliage: ants at extrafloral nectaries of Hibiscus pernambucensis in sandy and mangrove forests. Entomologia Experimentalis Et Applicata, 107, 125‐131. Costanza R, Darge R, Degroot R et al. (1997). The value of the world's ecosystem services and natural capital. Nature, 387, 253‐260. Danielsen F, Heegaard M (1995). Impact of logging and plantation development on species diversity: a case study from Sumatra. In: Sandbukt, O (ed). Management of tropical forests: towards an integrated perspective. University of Oslo, pp. 73‐92. Díaz S, Lavorel S, de Bello F, Quétier F, Grigulis K, Robson M (2007). Incorporating plant functional diversity effects in ecosystem service assessments. Proceedings of the National Academy of Sciences of the United States of America, 104, 20684‐20689. Didham RK, Hammond PM, Lawton JH, Eggleton P, Stork NE (1998). Beetle species responses to tropical forest fragmentation. Ecological Monographs, 68, 295‐323. Fahrig L (2003). Effects of habitat fragmentation on biodiversity. Annual Review of Ecology Evolution and Systematics, 34, 487‐515. 18 Chapter 1 General Introduction FAOSTAT (2009). FAO Statistical Databases. Available from URL http://faostat.f ao.org/ Fontaine C, Dajoz I, Meriguet J, Loreau M (2006). Functional diversity of plant‐pollinator interaction webs enhances the persistence of plant communities. Plos Biology, 4, 129‐135. Gallai N, Salles JM, Settele J, Vaissiere BE (2009). Economic valuation of the vulnerability of world agriculture confronted with pollinator decline. Ecological Economics, 68, 810‐821. Goehring DM, Daily GC, Şekercioğlu CH (2002). Distribution of ground‐dwelling arthropods in a tropical countryside. Journal of Insect Conservation, 6, 83‐91. Grimbacher PS, Catterall CP, Kitching RL (2006). Beetle species’ responses suggest that microclimate mediates fragmentation effects in tropical Australian rainforest. Austral Ecology, 31, 458‐470. Heil M, Mckey D (2003). Protective ant‐plant interactions as model systems in ecological and evolutionary research. Annual Review of Ecology Evolution and Systematics, 34, 425‐453. Herrera CM (2002). Seed dispersal by vertebrates. In: Herrera CM and Pellmyr O, ed. Plant–animal Interactions: An Evolutionary Approach. Blackwell Science, Oxford, pp. 185–210. Hu HB, Liu WJ, Cao M (2008). Impact of land use and land cover changes on ecosystem services in Menglun, Xishuangbanna, Southwest China. Environmental Monitoring and Assessment, 146, 147‐156. Jordano P (1992). Fruits and frugivory. In: Fenner M, ed. Seeds: The Ecology of Regeneration in Plant Communities. CAB International, Wallingford, UK, pp. 105‐156. Koh LP, Sodhi, NS (2010). Conserving Southeast Asia's imperiled biodiversity: scientific, management, and policy challenges. Biodiversity and Conservation, 19, 913‐917. Lewis OT (2009). Biodiversity change and ecosystem function in tropical forests. Basic and Applied Ecology, 10, 97‐102. Liu WJ, Hu HB, Ma YX, Li HM (2006). Environmental and socioeconomic impacts of increasing rubber plantations in Menglun township, southwest China. Mountain Research and Development, 26, 245‐253. López‐Bao JV, González‐Varo JP (2011). Frugivory and spatial patterns of seed deposition by carnivorous mammals in anthropogenic landscapes: a multi‐scale approach. PLoS One, 6, e14569. Lü XT, Yin JX, Tang JW (2010). Structure, tree species diversity and composition of tropical seasonal rainforests in xishuangbanna, south‐west china. Journal of Tropical Forest 19
Chapter 1 General Introduction Science, 22, 260‐270. Mcgarigal K, Cushman SA (2002). Comparative evaluation of experimental approaches to the study of habitat fragmentation effects. Ecological Applications, 12, 335‐345. Moran C, Catterall CP, Kanowski J (2009). Reduced dispersal of native plant species as a consequence of the reduced abundance of frugivore species in fragmented rainforest. Biological Conservation, 142, 541‐552. Morris RJ (2010). Anthropogenic impacts on tropical forest biodiversity: a network structure and ecosystem functioning perspective. Philosophical Transactions of the Royal Society B‐Biological Sciences, 365, 3709‐3718. Muscarella R, Fleming TH (2007). The role of frugivorous bats in tropical forest succession. Biological Reviews, 82, 573‐590. Peh KSH, de Jong J, Sodhi NS, Lim SLH, Yap CAM (2005). Lowland rainforest avifauna and human disturbance: persistence of primar forest birds in selectively logged forests and mixed‐rural habitats of southern Peninsular Malaysia. Biological Conservation, 123, 489‐505. Perfecto I, Vandermeer J (2008). Biodiversity conservation in tropical agroecosystems ‐ A new conservation paradigm. Year in Ecology and Conservation Biology 2008. Oxford: Blackwell Publishing. Reagan DP, Waide RB (1996). The Food Web of a Tropical Rain Forest. University of Chicago Press, Chicago. Ricketts TH, Regetz J, Steffan‐Dewenter I et al. (2008). Landscape effects on crop pollination services: are there general patterns? Ecology Letters, 11, 499‐515. Rosumek FB, Silveira FAO, Neves FD et al. (2009). Ants on plants: a meta‐analysis of the role of ants as plant biotic defenses. Oecologia, 160, 537‐549. Roubik DW (1995). Pollination of cultivated plants in the tropics. FAO Bulletin 118. Sodhi NS, Posa MRC, Lee TM, Bickford D, Koh LP, Brook BW (2010). The state and conservation of Southeast Asian biodiversity. Biodiversity and Conservation, 19, 317‐328. Steffan‐Dewenter I, Westphal C (2008). The interplay of pollinator diversity, pollination services and landscape change. Journal of Applied Ecology, 45, 737‐741. Tscharntke T, Sekercioglu CH, Dietsch TV, Sodhi NS, Hoehn P, Tylianakis JM (2008). Landscape constraints on functional diversity of birds and insects in tropical agroecosystems. Ecology, 89, 944‐951. Tylianakis JM, Laliberté E, Nielsen A, Bascompte J (2010). Conservation of species interaction networks. Biological Conservation, 143, 2270‐2279. 20 Chapter 1 General Introduction Vander Wall SB, Kuhn KM, Beck MJ (2005b). Seed removal, seed predation, and secondary dispersal. Ecology, 86, 801‐806. Vander Wall SB, Kuhn KM, Gworek JR (2005a). Two‐phase seed dispersal: linking the effects of frugivorous birds and seed‐caching rodents. Oecologia, 145, 282‐287. Wang BC, Smith TB (2002). Closing the seed dispersal loop. Trends in Ecology & Evolution, 17, 379‐385. Wright SJ (2005). Tropical forests in a changing environment. Trends in Ecology & Evolution, 20, 553‐560. Wu ZL, Liu HM, Liu LY (2001). Rubber cultivation and sustainable development in Xishuangbanna, China. International Journal of Sustainable Development and World Ecology, 8, 337‐345. Zhang JH, Cao M (1995). Tropical forest vegetation of xishuangbanna, sw china and its secondary changes, with special reference to some problems in local nature conservation. Biological Conservation, 73, 229‐238. Zhang KY (1986). The influence of deforestation of tropical rainforest on local climate and disaster in Xishuangbanna region of China. Climatol. Notes, 35, 223‐236. Zhu H (2008). The tropical flora of southern yunnan, china, and its biogeographic affinities. Annals of the Missouri Botanical Garden, 95, 661‐680. Zhu H, Cao M, Hu HB (2006). Geological history, flora, and vegetation of Xishuangbanna, southern Yunnan, China. Biotropica, 38, 310‐317. 21
Chapter 1 22 General Introduction Chapter 2 Carabid beetle diversity Chapter 2 Carabid beetle communities and species distribution in a changing tropical landscape (southern Yunnan, China) Meng, L.‐Z., Martin, K., Weigel, A., Liu, J.‐X (2011): Impact of rubber plantation on carabid beetle communities and species distribution in a changing tropical landscape (southern Yunnan, China). Journal of Insect Conservation, DOI:10.1007/s10841‐011‐9428‐1. 23
Chapter 2 Carabid beetle diversity Abstract Carabid beetles (Coleoptera: Carabidae) have widely been used to assess biodiversity values of different habitats in cultivated landscapes, but rarely in the humid tropics. This study aimed to investigate effects of land use change on the carabid assemblages in a tributary valley of the Mekong River in tropical southern Yunnan, China. The study area includes habitats of traditional land use systems (rice production and shifting cultivation successions) and was dominated by natural forests until about 30 years ago. Since then, large areas of forest have been, and still are, successively transformed into commercial rubber monoculture plantations. In total, 102 species of Carabidae (including Cicindelinae) were recorded from 13 sites over different seasons, using pitfall traps, Malaise traps and aerial collectors in trees. Cluster analysis and indicator species analysis showed that three types of habitat (rice field fallows, early natural successions and natural forest) possess a degree of uniqueness in species composition. Non‐metric multidimensional scaling revealed that the environmental factors explaining 80% of the total variation in carabid assemblage composition are the degree of vegetational openness of a habitat and its plant species diversity. Rice field fallows had significantly higher numbers of species and individuals than any other type of habitat and are probably dominated by species originating from other regions. Carabid assemblages of young rubber plantations (5 and 8 years) were quantitatively similar to those of forests, but without species of significant indicator value. With increasing plantation age (20 and 40 years), the number of carabid species decreased. Increasing age and a further spatial expansion of rubber plantations at the expense of forest areas will have negative impacts on the native forest carabid assemblages with strongest effects on forest specialists and rare species. Keywords: Biodiversity; Ground beetles; Land use change; Rubber plantation; Succession; Tropical forest 24 Chapter 2 Carabid beetle diversity 2.1 Introduction Carabid beetles (Coleoptera: Carabidae) have been widely used to assess biodiversity values of different habitats in cultivated landscapes usually composed of heterogeneous mosaics of various land uses. Most species are carnivorous and actively hunt for any small invertebrate prey they can overpower and some species are polyphagous which feed the small seeds of herbaceous plants (Honek et al. 2003). Specific aspects of carabid beetle diversity and distribution include e.g. spatio‐temporal changes across gradients of vegetation, land use intensification or disturbance (Aviron et al. 2005; Gobbi & Fontaneto 2008; Gu et al. 2008, Roughley et al. 2006; Yu et al. 2007), effects of habitat fragmentation and isolation (Fujita et al. 2008; Wamser et al. 2010), and effects of land management and cropping systems (Eyre et al. 2009; O’Rourke et al. 2008), grasslands (Purtauf et al. 2004), forests and forest plantations (Fuller et al. 2008; Taboada et al. 2008; Yu et al. 2008). Conclusions drawn from those studies are that most carabid species can be subdivided in open land and forest species, with specialists and generalists in each group. Furthermore, increasing intensification and disturbance in cultivated landscapes tends to reduce forest specialist carabids and to homogenize the beetle assemblages between the habitat types. Nearly all studies were conducted in temperate or subtropical regions, where the original vegetation has disappeared or been strongly modified in the course of a usually long history of land cultivation. Very little research has been conducted on the effects of land use change on carabid assemblages in relatively young cultivated landscapes of tropical rainforest regions (but see Goehring et al. 2002 and Gormley et al. 2007 for Costa Rica, and Rainio & Niemelä 2006 for Madagascar). In this study, we investigate the carabid assemblages of the major land use types in a tropical landscape of southern Yunnan, China. The region is part of the ‘Indo‐Burma hotspot’, one of the 34 global hotspots exceptionally rich in biodiversity (Biodiversity Hotspots 2007). The study area represents a tributary valley of the Mekong River. There, traditional land use systems are irrigated rice fields along the river courses and shifting cultivation systems on the slopes, but the largest proportion of the land area was covered with primary and secondary forest until about 30 years ago. Since then, large areas of forest have been, and still are, successively transformed into commercial rubber monoculture plantations of different age. This pattern is representative of the development of tropical southern Yunnan. Detailed data from a 25
Chapter 2 Carabid beetle diversity typical subregion showed that between 1988 and 2006, rubber plantations increased from 12% of the total land cover to 46%, whereas forested areas dropped from 49% to 28% (Hu et al. 2008). Tropical seasonal rainforest was the type of land most affected by the expansion of rubber plantations (Li et al. 2007). 2.2 Materials and Methods 2.2.1 Study area and sampling sites The study was conducted in the Naban river valley (ca. 11,000 ha), within the area of the Naban River Watershed National Nature Reserve (NRWNNR) in Xishuangbanna, southern Yunnan province, south‐west China (22° 10’ N and 100° 38’ E). The region represents the northernmost part of the humid tropics in Asia with a climate influenced by Monsoon and three distinct seasons: cool‐dry (October‐January, with the lowest monthly temperature of 15°C in December), hot‐dry (February‐April, with the highest monthly temperature of 25 °C in April) and a rainy season (May‐September) with most of the mean annual precipitation of almost 1600 mm. The natural vegetation of the study region is tropical rainforest, falling into different types of evergreen and seasonal forests related to topography and elevation (Cao et al. 2006; Lü et al 2010). Secondary and primary forest sites and fragments are still widespread in the study area, but most cultivated land is covered by plantations of rubber (Hevea brasiliensis). The remaining land use types are mainly rice fields in the valley bottoms and other crops around the small villages, and grassland and shrubland successions. A map of the current land use (Fig. 2.1) was derived from IKONOS satellite imagery (acquired on November 16 and December 2, 2007) via supervised classification using ERDAS Imagine software (Berkhoff et al. 2009). To represent the most typical land‐use types of this landscape, 13 sites including forest plots, rubber plantations and different types of open land (Table 2.1) were selected for recording the Carabidae, including the subfamily of tiger beetles (Cicindelinae). 26 Chapter 2 Carabid beetle diversity Fig. 2.1 Major land use types along the Naban River valley within the boundaries of the Naban River Watershed National Nature Reserve, and locations of the study sites according to Table 1. Table2.1 Habitat characteristics of the 13 sampling sites and the environmental variables used in the NMS analysis: Total number of vascular plant species (Plant) and species numbers of different life forms; percentage of ground vegetation cover (GroCov) and canopy cover (CanCov); the successional age of the site or age of the trees (SucAge), vegetation height (VegHei) and four categories of land use. GroCov CanCov SucAge VegHei Land
No. of plant species Study site (Code) (%) (years) (m) use
Plan Grass Forb Liana Shrub Tree (%) Forest (MD‐FO) Forest (NB‐FO) 91 7 119 2 14 17 10
15
26 34 34
51
50 68 95 90 80 70 30.3 33.4 3 3 Forest (AM‐FO) 93 10 14 10
24 35
73 95 80 21.5 3 Forest (GMS‐FO) 96 18 17 10
18 33
75 85 60 28.5 3 Rubber (MD‐RU) 53 16 13 5 6 13
30 75 5 10.1 2 Rubber (NB‐RU) 49 12 21 2 5 9
21 87 8 11.7 2 Rubber (AM‐RU) 45 10 10 4 11 10
12 94 20 20 2 Rubber (SYD‐RU) 18 5 7 1 3 2
5 95 40 30 2 Clear fell (NB‐OP) 45 8 18 3 12 6
85 1 2 2.5 4 Grassland (AM‐OP) 57 18 18 3 12 6
75 1 5 1.2 4 Shrubland (GMS‐OP) 61 20 17 3 14 7
90 1 25 1.2 4 Rice fallow (MD‐FA) 53 16 35 0 1 1
95 1 1 0.7 1 Rice fallow (GMS‐FA) 60 22 38 0 0 1
97 1 1 0.6 1 27
Chapter 2 Carabid beetle diversity 2.2.2 Field methods Beetle sampling was carried out by using a combined trap system including pitfall traps and Malaise traps (Townes 1962) at all sites, and aerial collectors in the canopy area of trees in forests and in rubber plantations. Pitfall traps were plastic pots with a diameter of 8.5 cm and a depth of 13 cm buried flush to the soil surface, one third filled with 10% formalin solution. At each site, five pitfall traps were arranged at a distance of ca. 3 m from each other around a Malaise trap. Aerial collectors were constructed of two pieces of transparent plastic plates (50 x 30 cm, height x width) which were arranged crosswise and fixed upon a red plastic bowl of 30 cm in diameter. These traps were installed on canopy tree branches using ropes. The collecting bottles of the Malaise traps and the bowls of the aerial collectors were filled with a mixture of liquid of blue colored anti‐freeze (ethanol‐glycol). Traps were installed in different seasonal periods covering (a) the beginning of the rainy season (May‐July 2008), (b) the beginning of the cool‐dry season (September‐November 2008), and (c) the transition period from the hot‐dry to the rainy season (March‐June 2009). All traps were emptied every 10 days during the collecting periods (with few exceptions where traps were destroyed or collection was impossible due to heavy rains). The beetle specimens were preserved in 70% ethanol for further identification to the lowest possible taxonomic level. Data analyses are based on numbers of species and individuals combined from all traps and trap types per site and the total counts from all collecting periods. Voucher specimen of the collected beetles are kept at the National Zoological Museum of China, Institute of Zoology, CAS, Beijing. Vascular plant species inventories from each of the 13 trap sites were recorded in March 2009. At the four natural forest sites, four 20 x 20 m2 plots were established around the trap locations to record the numbers of tree and liana species. The other plant species (representing the groundcover vegetation < 2 m) in the forest plots were recorded from four 5 x 5 m2 subplots within each of the four large plots. Total species numbers of the four small and the four large plots were used for further calculations. In the other sites, records from all four 5 x 5 m2 plots per site provided the plant species numbers used for calculations. Voucher specimens of the recorded and identified plant species were deposited at the Herbarium of the Xishuangbanna Tropical Botanical Garden (XTBG), CAS, Yunnan, China. 28 Chapter 2 Carabid beetle diversity 2.2.3 Data analyses Cluster analyses were performed to identify quantitatively similar groups of carabid assemblages among the different sites and habitats. We used the Morisita index of similarity (Morisita 1959), being recommended as the best overall similarity measure in community analysis (Wolda 1981). Among the algorithms for hierarchical clustering, we selected the unweighted pair‐group method using averages (UPGMA) which is conventionally used in ecology (James & McCulloch 1990; Wolda 1981). One‐way ANOSIM (analysis of similarities) global tests were then applied to test for differences between subgroups produced by the cluster analysis. Differences in total species numbers and abundances of carabids between the subgroups were compared by the Mann‐Whitney U‐test using Minitab 15.0 software (Minitab Inc. State College PA, USA). This non‐parametric test was applied because the assumptions of homogeneity of variances and normality (tested with the Shapiro‐Wilk normality test) were not met according to Zar (1996). Non‐metric multidimensional scaling (NMS; Kruskal 1964) using the Bray–Curtis index for abundance data was applied to display and test for differences in carabid assemblage composition across the habitat types. This ordination scores was performed with PC‐ORD software (McCune & Mefford 2006) with the following parameters employed in the NMS procedure: Sorensen distance measure; a maximum number of 500 iterations; random starting coordinates; 100 runs with real data; step down in dimensionality (initial step length = 0.2); 100 runs with randomized data. A total of 11 vegetation and land use variables after log transformed were included in the NMS analysis to test their effects on the carabid assemblages. Six variables refer to plant species richness, including the total number of vascular plant species per site and the species numbers of different life forms, i.e. grasses, forbs (non‐woody plants other than grasses), trees and lianas. The remaining variables are the degree of tree canopy cover and the degree of ground vegetation cover, the maximum vegetation height, the successional age of the study site (years after establishment of the present vegetation or age of the trees), and the discrimination between four categories of land use type, represented by rice field fallows, early successions (forest clear fell, grassland, scanty shrubland), rubber plantation and natural forest (Table 2.1). Correlations between the ordination and the environmental variables were calculated with the Pearson coefficient. Indicator Species Analysis based on the combined values of relative abundance and relative 29
Chapter 2 Carabid beetle diversity frequency of species (McCune & Grace 2002) was used to identify carabid species affiliated with specific land use types. The indicator value of each of the recorded species was calculated with PC‐ORD software (McCune & Mefford 2006) using 4999 runs in a Monte Carlo test considering values at P<0.05. 2.3 Results A total number of 102 carabid species (including 16 species of Cicindelinae) and 1649 individuals were recorded across all study sites and recording periods (Appendix 2.1). The cluster analysis dendrogram based on the quantitative similarities of the carabid assemblages is shown in Fig. 2.2. As supported by global one‐way ANOSIM tests, meaningful differences between assemblages occurred at a similarity of 0.5 and generated 4 subgroups (global R = 0.94; P < 0.0001). Subgroup 1 is represented by three sites, including the oldest rubber plantation (40 years), a forest clear fell site and an adjacent forest fragment. Subgroup 2 includes the three younger rubber plantations (5, 8 and 20 years) and three forest sites. Subgroup 3 represents two early succession sites (grassland and shrubland), and subgroup 4 is formed by the two rice 1
S
YD
N RU
BO
N P
BF
A O
M
-R
M U
D
-F
N O
BR
A U
M
-F
G O
M
SM FO
D
-R
A U
M
-O
G P
M
SG OP
M
SM FA
D
-F
A
field fallows. Similarity (Morisita-index)
0.8
0.6
1
2
3
4
0.4
0.2
Fig. 2.2 Quantitative similarity cluster analysis of the carabid communities at the sampling localities (site codes see Table 2.1), generated from the Morisita‐index using UPGMA. The dendrogram shows 4 subgroups at similarity levels >0.5, indicated by the dashed line. 30 Chapter 2 Carabid beetle diversity 20
80
a
Number of carabid individuals
Number of carabid species
15
10
b
5
a
a
a
b
60
40
20
b
a
a
a
0
0
1
2
3
4
1
2
3
4
Fig. 2.3 Number of species (a) and number of individuals (b) of carabids in the four subgroups produced by cluster analysis (1–4, see Figure2.2). Box and whisker plots illustrate the 5th, 25th, 50th (median), 75th, and 95th percentiles, and the means as the dashed line. Different letters indicate significant differences (P<0.05; Mann‐Whitney U‐Test). Fig. 2.3 shows the differences in (a) numbers of species and (b) numbers of individuals of carabids between the four subgroups generated by cluster analysis. Subgroups 1, 2 and 3 have the lowest numbers of both species and individual numbers which were not significantly different from each other. Subgroup 4 had significantly higher numbers in all comparisons (P<0.05; Mann‐Whitney U‐test). Two axes were recommended by the NMS ordination of carabid assemblages (Fig.2.4) which together explain 80% of the total variance in the data set (Axis 1 = 48%, Axis 2 =32%) at a final stress of 13.5. The final instability was 0.00001 with 500 iterations. All environmental variables except the number of shrub species and land use category showed significant effects at P<0.05 (Pearson correlation coefficient). Axis 1 represents a gradient along the successional development with canopy‐covered habitats (forests and rubber plantations) on the negative side and open land (rice field fallows and early successions) on the positive side. Axis 2 shows a positive correlation with total plant species richness per site. Thus, the environmental factors explaining totally 80% of the variation in carabid assemblage composition are the degree of veg‐‐etational openness of a habitat and its plant species diversity. A total of 11 ground beetle species with indicator values at P<0.05 were identified (see Appendix for species indication). Of these, 7 species were specialists of rice field fallows including Acupalpus sp. 2, Aephnidius sp., Chlaenius laetiusculus, Harpalini Gen. sp. 4, Platymetopus sp., Stenolophus quinquepustulatus and Stenolophus sp. 3, two species were specialists of early successions (both representing Cicindelinae) including Calochroa interruptofasciata and Lophyra lineifrons, and two species were natural forest specialists including Orthogonius sp. 2 and Pheropsophus cf. beckeri. No species were significantly affiliated with rubber plantations.
31
Chapter 2 Carabid beetle diversity Fig. 2.4 NMS ordination of carabid communities at the 13 sample sites: (▲) rubber plantation; (▽) grassland and shrubland; (◆) rice field fallow; (●) forest. Variables with a Pearson correlation coefficient at P<0.05 are shown. Cumulative variation in the original dataset explained by ordination is 80% (Axis 1 = 47%, Axis 2 = 33%, Final stress = 13.5, Final instability = 0.00001). 2.4 Discussion The overall results regarding the carabid assemblage patterns of the valley landscape indicates that three land‐use types (rice field fallows, early successions and natural forest) possess a degree of uniqueness in species composition, each characterized by species with significant indicator values. Only 11 habitat specialist species were identified out of totally 102 recorded species. The numbers of specialists might indeed be higher if we used the additional analysis of estimated species number through Chao estimator, because most of the species were only represented by few individuals which were not enough to show statistically significant results. For example, from the 59 species recorded from forests, 14 species with numbers between one and 6 individuals were only found there. Although they were not denoted as indicator species, there might be forest specialists among them which are 32 Chapter 2 Carabid beetle diversity merely rare. The highest numbers of total species and individuals were recorded from the rice field fallows (subgroup 4 in the cluster analysis, Fig. 2.2 and 2.3). Since the natural vegetation in the study region is tropical forest, it can be concluded that the 7 specialist species and certainly others originate from naturally open habitat types of other regions. This can also be assumed for the carabid species typical of the grassland and shrubland in the study area. These habitats are characterized by the highest abundances of most of the ground‐dwelling Cicindelinae in an assemblage group distinct from other habitats (subgroup 3 in the cluster analysis, Fig. 2.2). The youngest rubber plantations with open canopy (5 and 8 years old) contain carabid species which were also recorded from other habitat types, but their quantitative assemblage composition show highest similarity to that of the forests (subgroup 2 in the cluster analysis, Fig. 2.2). The 20 year old rubber plantation with a closed canopy belongs to the same subgroup but has the greatest community distance from the other sites. The 40 year old rubber plantation falls into an assemblage group characterized by highly degraded or disturbed habitats of original forest, along with a forest clear fell and a small forest fragment (subgroup 1 in the cluster analysis, Fig. 2.2). Most carabid species of this group are ubiquitous species which occur in both open land and closed canopy habitats. It might be expected that vegetation openness to be important for ground surface active predatory carabids, while plant species diversity might be expected to be important for seed‐feeding Harpalines species. Genus including Acupalpus and Chlaenius species like wet and marshy locations and hence associated more with habitat of rice fallow than forest. Harpalus sp. are very small seed‐feeders relying strongly on the seed supply from different weeds and therefore favour openness of habitat and were mostly found in the pastures (Gu et al. 2008). Pheropsophus species are predators and known to be forest specialists to predate the eggs of mole cricket. Overall, the present land use types of the study area show a dynamic pattern of closed and open habitats with distinct and characteristic carabid assemblages, except those of the rubber plantations. Most of the rubber plantations in the study area are less than 20 years old, but will reach an age of about 40 years before the latex productivity declines. Then, the plantations will be clear felled for starting a new plantation cycle with young trees. Although the carabid assemblages of young rubber plantations are similar to those of the forest sites, the results also indicate that rubber plantations do not provide alternative habitats for the species indicated as 33
Chapter 2 Carabid beetle diversity being forest specialists. Young plantations are temporarily suitable for many carabid species, probably because they provide a ground cover vegetation of relatively high density and diversity in addition to the presence of a canopy. However, with increasing plantation age and the close of the canopy, the ground cover vegetation is shaded out and disappears, as shown by the 40 year old rubber plantation with an extremely low plant species richness (Table 2.1). This may explain that the lowest total carabid species richness of all habitat categories was recorded from the 20‐40 year old rubber plantations with closed canopies (25 species) compared to the 5‐8 year old plantations with open canopies (34 species) and natural forest (59 species, see Appendix 2.1). The question of how non‐indigenous tree plantations, such as conifers or Eucalyptus, affect the diversity of and composition of carabid assemblages has been addressed in other landscapes. It was generally confirmed that carabid assemblage composition changes with increasing plantation age, which is usually related to a reduction of open land species (Jukes et al. 2001; Karen et al. 2008; Pawson et al. 2009). However, no overall conclusions can be drawn on the value of mature tree plantation stands for the conservation of carabids originating from natural forests. There are studies showing a reduced carabid species richness in plantations compared to forests (da Silva et al. 2008; Fahy & Gormally 1998; Magura et al. 2003) or a lower number of forest specialist species (Fuller et al. 2008). On the other hand, it was shown that carabid beetle assemblages of mature plantations do not significantly differ from those of natural or semi‐natural forest (Elek et al. 2010; Karen et al. 2008; Martinez et al. 2009), indicating that the establishment of tree plantations will not lead to a reduced carabid diversity at the landscape scale. Various environmental factors may account for such differences. Although the degree of canopy cover proved to be an important factor of the habitat suitability for forest carabid species, the structure and diversity of the ground cover vegetation is also closely related to the species richness and abundance of carabids, as shown in the present study and in others (Karen et al. 2008; Oxbrough et al. 2010; Pawson et al. 2009; Yu et al. 2008). However, it is difficult to distinguish between the effects of tree species composition and spatial heterogeneity on the species richness of the ground vegetation and ultimately the carabid species richness in different forest and plantation types (Taboada et al. 2010). Another factor influencing the differences in carabid composition between forest and plantations in a given landscape can be the habitat quality of forest plots, indicated by the degree of disturbance, fragmentation or size. For example, Fujita et 34 Chapter 2 Carabid beetle diversity al. (2008) found that the richness of forest carabid species markedly decreased with the reduction of area of forest fragments. The poor habitat quality of patchy native forest remnants enclosed by plantations was assumed to be the main factor explaining the high similarity in carabid diversity of the two habitat types in the study of Martinez et al. (2009). 2.5 Conclusions To assess the effects of rubber plantations on the diversity and distribution of carabid species in the study region, temporal and spatial changes need to be considered. Currently most of the rubber plantations represent young stages of development which provide a transient habitat for forest generalists from the neighboring forest areas and for ubiquitous species. With increasing plantation age, however, habitat quality decreases for all species. A further expansion of rubber cultivation will finally result in large areas of mature rubber plantations with negative impacts on the native forest species populations and strongest effects on forest specialists and rare species. Because rubber cultivation largely proceeds at the expense of forest areas and not of cultivated land, the carabid assemblages of open habitats will be less affected by this scenario. It is therefore essential to protect natural forest areas as a pool for forest carabid species. Effects of rubber plantations could then be mitigated by a management of the rotation cycles that allows the steady presence of young plantation stages with a ground cover vegetation that can serve as a temporary habitat for carabid species. 2.6 References Aviron S, Burel F, Baudry J, Schermann N (2005). Carabid assemblages in agricultural landscapes: impacts of habitat features, landscape context at different spatial scales and farming intensity. Agriculture Ecosystems & Environment, 108, 205‐217. Berkhoff K, Cotter M, Herrmann S, Sauerborn J eds(2009). Using remote sensing data as basic information for applied land use change modelling. Proceedings of the ERSEC International Conference 2008, Sustainable Land Use and Water Management; October 2008, Beijing, China. pp 36‐45. Biodiversity Hotspots (2007). Available from URL http://www.biodiversityhotsp ots.org/ 35
Chapter 2 Carabid beetle diversity Cao M, Zou X, Warren M, Zhu H (2006). Tropical forests of Xishuangbanna, China. Biotropica, 38, 306‐309. da Silva PM, Aguiar CAS, Niemelä J, Sousa JP, Serrano ARM (2008). Diversity patterns of ground‐beetles (Coleoptera: Carabidae) along a gradient of land‐use disturbance. Agriculture Ecosystems & Environment, 124, 270‐274. Elek Z, Dauffy‐Richard E, Gosselin F (2010). Carabid species responses to hybrid poplar plantations in floodplains in France. Forest Ecology and Management, 260, 1446‐1455. Eyre MD, Labanowska‐Bury D, Avayanos JG, White R, Leifert C (2009). Ground beetles (Coleoptera: Carabidae) in an intensively managed vegetable crop landscape in eastern England. Agriculture Ecosystems & Environment, 131, 340‐346. Fahy O, Gormally M (1998). A comparison of plant and carabid beetle communities in an Irish oak woodland with a nearby conifer plantation and clearfelled site. Forest Ecology and Management, 110, 263‐273. Fujita A, Maeto K, Kagawa Y, Ito N (2008). Effects of forest fragmentation on species richness and composition of ground beetles (Coleoptera: Carabidae and Brachinidae) in urban landscapes. Entomological Science, 11, 39‐48. Fuller RJ, Oliver TH, Leather SR (2008). Forest management effects on carabid beetle communities in coniferous and broadleaved forests: implications for conservation. Insect Conservation and Diversity, 1, 242‐252. Gobbi M, Fontaneto D (2008). Biodiversity of ground beetles (Coleoptera: Carabidae) in different habitats of the Italian Po lowland. Agriculture Ecosystems & Environment, 127, 273‐276. Goehring DM, Daily GC, Şekercioğlu CH (2002). Distribution of ground‐dwelling arthropods in a tropical countryside. Journal of Insect Conservation, 6, 83‐91. Gormley LHL, Furley PA, Watt AD (2007). Distribution of ground‐dwelling beetles in fragmented tropical habitats. Journal of Insect Conservation, 11, 131‐139. Gu W, Sang W, Liang H, Axmacher JC (2008). Effects of Crofton weed Ageratina adenophora on assemblages of Carabidae (Coleoptera) in the Yunnan Province, South China. Agriculture Ecosystems & Environment, 124, 173‐178. Honek A, Martinkova Z, Jarosik V (2003). Ground beetles (Carabidae) as seed predators. European Journal of Entomology, 100, 531‐544 Hu HB, Liu WJ, Cao M (2008). Impact of land use and land cover changes on ecosystem services in Menglun, Xishuangbanna, Southwest China. Environmental Monitoring and Assessment, 146, 147‐156. 36 Chapter 2 Carabid beetle diversity James FC, McCulloch CE (1990). Multivariate‐analysis in ecology and systematics ‐ panacea or pandora box. Annual Review of Ecology and Systematics, 21, 129‐166. Jukes MR, Peace AJ, Ferris R (2001). Carabid beetle communities associated with coniferous plantations in Britain: the influence of site, ground vegetation and stand structure. Forest Ecology and Management, 148, 271‐286. Karen M, O’Halloran J, Breen J, Giller P, Pithon J, Kelly T (2008). Distribution and composition of carabid beetle (Coleoptera, Carabidae) communities across the plantation forest cycle ‐ Implications for management. Forest Ecology and Management, 256, 624‐632. Kruskal JB (1964). Nonmetric multidimensional scaling: a numerical method. Psychometrika, 29, 115‐129. Li H, Aide TM, Ma Y, Liu W, Cao M (2007). Demand for rubber is causing the loss of high diversity rainforest in SW China. Biodiversity and Conservation, 16, 1731‐1745. Lü XT, Yin JX, Tang JW (2010). Structure, tree species diversity and composition of tropical seasonal rainforests in Xishuangbanna, south‐west China. Journal of Tropical Forest Science, 22, 260‐270. Magura T, Tóthmérész B, Elek Z (2003). Diversity and composition of carabids during a forestry cycle. Biodiversity and Conservation, 12, 73‐85. Martinez A, Iturrondobeitia JC, Goldarazena A (2009). Effects of some ecological variables on carabid communities in native and non native forests in the Ibaizabal basin (Basque Country: Spain). Annals of Forest Science, 66, 304‐314. McCune B, Grace J (2002). Analysis of ecological communities. MjM Publishers, Gleneden Beach, Oregon. USA. McCune B, Mefford MJ (2006). PC‐ORD. Multivariate analysis of ecological data. Version 5.10. MjM Software, Gleneden Beach, Oregon, USA. Morisita M (1959). Measuring of the dispersion and analysis of distribution patterns. Mem Fac Sci Kyushu Univ. Ser. E. (Biol), 2, 215‐235. O’Rourke ME, Liebman M, Rice ME (2008). Ground beetle (Coleoptera: Carabidae) assemblages in conventional and diversified crop rotation systems. Environmental Entomology, 37, 121‐130. Oxbrough A, Irwin S, Kelly TC, O’Halloran J (2010). Ground‐dwelling invertebrates in reforested conifer plantations. Forest Ecology and Management, 259, 2111‐2121. Pawson SM, Brockerhoff EG, Didham RK (2009). Native forest generalists dominate carabid assemblages along a stand age chronosequence in an exotic Pinus radiata plantation. Forest Ecology and Management, 258, 108‐116. 37
Chapter 2 Carabid beetle diversity Purtauf T, Dauber J, Wolters V (2004). Carabid communities in the spatio‐temporal mosaic of a rural landscape. Landscape and Urban Planning, 67, 185‐193. Rainio J, Niemelä J (2006). Comparison of carabid beetle (Coleoptera: Carabidae) occurrence in rain forest and human‐modified sites in south‐eastern Madagascar. Journal of Insect Conservation, 10, 219‐228. Roughley RE, Pollock DA, Wade DJ (2006). Biodiversity of ground beetles (Coleoptera: Carabidae) and spiders (Araneae) across a tallgrass prairie‐Aspen forest ecotone in southern Manitoba. Canadian Entomologist, 138, 545‐567. Taboada A, Kotze DJ, Tárrega R, Salgado JM (2008). Carabids of differently aged reforested pinewoods and a natural pine forest in a historically modified landscape. Basic and Applied Ecology, 9, 161‐171. Taboada A, Tarrega R, Calvo L, Marcos E, Marcos JA, Salgado JM (2010). Plant and carabid beetle species diversity in relation to forest type and structural heterogeneity. European Journal of Forest Research, 129, 31‐45. Townes H (1962). Design for a Malaise trap. Proceedings of the Entomological Society of Washington, 64, 253‐262. Wamser S, Diekötter T, Boldt L, Wolters V, Dauber J (2010). (early view). Trait‐specific effects of habitat isolation on carabid species richness and community composition in managed grasslands. Insect Conservation and Diversity, DOI: 10.1111/j.1752‐4598.2010.00110.x . Wolda H (1981). Similarity indexes, sample‐size and diversity. Oecologia, 50, 296‐302. Yu XD, Luo TH, Zhou HZ, Yang J (2007). Distribution of carabid beetles (Coleoptera: Carabidae) across a forest‐grassland ecotone in southwestern China. Environmental Entomology, 36, 348‐355. Yu XD, Luo TH, Zhou HZ (2008). Distribution of carabid beetles among 40‐year‐old regenerating plantations and 100‐year‐old naturally regenerated forests in Southwestern China. Forest Ecology and Management, 255, 2617‐2625. Zar JH (1996). Biostatistical analysis, 3rd edn. Prentice Hall, Upper Saddle River, NJ. 38 Chapter 2 Carabid beetle diversity Appendix 2.1 List of carabid species and numbers of individuals recorded from the 13 study sites compiled by habitat type. The Cicindelinae are listed separately. Species with indicator species values at P<0.05 are underlined, and affiliation with habitat type is indicated by superscript letters (F= forest, R = rice field fallow, G = grassland and shrubland). Species Acupalpus sp. 1 Acupalpus sp. 2 R Acupalpus sp. 3 Aephnidius sp. R Agonini Gen. sp. 1 Agonini Gen. sp. 2 Agonini Gen. sp. 2 Amblystomus sp. 1 Amblystomus sp. 2 Amblystomus sp. 3 Amblops piceus Anisodactylus sp. Badister (Baudia) sp. Bembidion sp. Bradycellini Gen. sp. Calleida klapperichi Calleida sp. Callistoides caeruleiceps Callistomimus quadricolor Casnoidea indica Catascopus cf. facialis Catascopus sp. Chlaenius costiger Chlaenius circumdatus Chlaenius laetiusculus‐group R Chlaenius bimaculatus Chlaenius cambodiensis Clivina sp. Coptodera sp. Cymindis sp. Dicranoncus quadridens Dioryche sp. Dromiini Gen. sp. Dromius (Klepterus) sp. Dyschirius sp. Egadroma sp. Forest
0 0 0 0 1 2 8 1 0 0 2 0 0 3 0 6 0 0 0 1 7 1 2 0 0 1 0 0 3 1 0 1 1 0 0 10 Rubber
(5‐8y) 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 1 0 3 0 0 0 0 0 0 1 1 0 0 1 0 3 Rubber (20‐40y) 0 0 1 0 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 0 0 3 0 0 0 1 0 0 6 3 Grass, Rice shrub fallow
8 1 0 2 0 0 0 3 0 0 0 0 1 2 0 0 0 2 1 25 0 0 0 1 0 1 0 1 1 2 0 0 0 0 1 1 2 0 0 0 1 0 0 0 1 2 1 1 0 2 3 2 6 3 1 1 0 0 3 0 0 0 0 2 0 0 0 0 0 2 6 7 39
Chapter 2 Carabid beetle diversity Elaphropus poecilopterus Gen. sp. (Lebiini?) Harpalini Gen. sp. Harpalini Gen. sp. 1 Harpalini Gen. sp. 2 Harpalini Gen. sp. 3 Harpalini Gen. sp. 4 R Harpalus sp. 1 Harpalus sp. 2 Harpalus sp. 3 Holcoderus ? sp. Holosoma cf. opacum Lebia sp. 1 Lebia sp. 2 Lebiini Gen sp. Macrocheilus sp. 1 Macrocheilus sp. 2. Microlestes sp. Oodini Gen. sp. 1 Oodini Gen sp. 2 Oodini Gen sp. 3 Orthogonius sp. 1 Orthogonius sp. 2 F Orthogonius sp. 3 Orthogonius sp.4 Orthogonius sp. 5 Oxycentrus sp. Peliocyphas sp. Pentagonica sp. Perigonini gen. spec. Pheropsophus cf. beckeri F Pheropsophus cf. javanus Platymetopus sp. R Platynini Gen. sp. 1 Platynini Gen. sp. 2 Platynini Gen. sp. 3 Platynini Gen. sp. 4 Platynus sp. Pseudoophonus sp. Pterostichini Gen. sp. 1 Pterostichini Gen. sp. 2 Sphrodriini Gen. sp. Stenolophini Gen. sp. 1 Stenolophini Gen. sp. 2 40 0 1 0 0 0 3 2 0 0 2 1 3 1 1 4 0 2 1 1 3 0 92 119
24 41 8 0 4 15 2 24 0 0 0 0 1 4 3 0 1 0 4 0 0 0 0 2 0 0 1 2 1 0 1 0 0 0 1 0 2 1 0 0 1 0 25 27 3 11 3 1 0 16 0 0 0 0 2 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 0 0 0 0 0 0 1 0 0 3 7 7 7 1 2 1 3 0 0 0 2 0 0 0 1 1 0 0 0 0 0 0 0 0 0 1 0 7 2 20 2 22 0 1 1 0 0 1 0 1 0 0 1 15 43 5 7 5 7 0 11 0 0 0 0 0 1 0 1 0 1 0 0 0 0 0 1 0 1 1 3 3 6 10 1 15 0 0 0 0 1 0 0 4 0 1 0 50 11 9 7 1 0 0 9 0 1 3 3 3 2 1 0 1 5 0 0 0 1 1 Chapter 2 Carabid beetle diversity Stenolophus quinquepustulatusR Stenolophus sp. 2 Stenolophus sp. 3 R Syntomus sp. Trichotichnus sp. Trigonotma sp. Cicindelinae Callytron andersonii Calochroa elegantula Calochroa interruptofasciata G Calochroa salvazi Cosmodela virgula Cylindera holosericea Cylindera kaleea Cylindera spinolae Cylindera viduata Heptodonta eugenia Lophyra striolata Lophyra lineifrons G Myriochila sinica Neocollyris linearis Therates pseudorugifer Tricondyla mellyi Number of species per habitat type 1 0 5 1 5 0 2 16 3 1 0 0 1 18 0 5 0 6 0 2 1 1 59 1 0 1 0 2 0 0 7 0 0 0 0 1 4 0 1 0 1 0 0 0 0 34 0 0 1 0 1 0 0 29 0 0 0 0 0 4 0 0 0 4 0 0 0 0 25 2 1 0 0 2 1 6 46 30 12 2 3 18 24 3 5 1 108 5 2 0 0 53 139 1 31 1 2 1 7 10 0 6 1 0 24 20 0 1 0 12 0 2 0 1 64 41
Chapter 2 Carabid beetle diversity 42 Chapter 3 Hoverflies & Wild bees Chapter 3 Contrasting responses of hoverflies and wild bees to habitat structure and land use change in a tropical landscape (southern Yunnan, SW China) Meng, L.‐Z., K. Martin, Liu, J.‐X., F. Burger, Chen, J. (2012). Contrasting responses of hoverflies and wild bees to habitat structure and land use change in a tropical landscape (southern Yunnan, SW China). Insect Science, DOI: 10.1111/j.1744‐7917.2011.01481.x. 43
Chapter 3 Hoverflies & Wild bees Abstract How pollinating insects response to monoculture plantations mainly proceeded at the expense of natural forest areas is an outstanding and important issue in ecology and conservation biology with the pollination services declined around the world. In this study, species richness and distribution of hoverfly and wild bee communities were investigated in a changing tropical landscape in southern Yunnan, south‐west China by Malaise trap from 2008 to 2009 periodically. Species were recorded from the traditional land use types (natural forest, grassland, shrubland and rice field fallows), and from recently established rubber plantations of different ages. Hoverflies (total 53 species) were most common in young successional stages of vegetation including rice field fallow and shrubland. Species richness was highest in rice field fallows and lowest in forests and showed a highly significant relationship with the number of forb species and ground vegetation cover. In contrast, the highest richness of wild bees (total 44 species) was recorded from the natural forest sites, which showed a discrete bee community composition compared to the remaining habitat types. There was no significant relationship between the bee species richness and the environmental variables including the numbers of different plant life form, coverage of canopy and ground vegetation, successional age of vegetation and the land use type. At landscape scale, open land use systems including young rubber plantations were expected to increase the species richness of hoverflies, however might decrease wild bees diversity. The present land use change by rubber cultivation can be expected to have negative impacts on the native wild bee communities. Keywords: Apidae; insects; pollinators; rubber plantations; Syrphidae; tropical forest 44 Chapter 3 Hoverflies & Wild bees 3.1 Introduction Pollinators are a key factor for the biodiversity of wild flowering plants, and they provide an economically important ecosystem service to global crop production. More than one third (35%) of the world food production depends on animal pollination (Klein et al. 2007). About 70% of 1330 tropical crops benefit from pollination by animals, including naturally occurring insects such as wild bees, butterflies, moths, and hoverflies (Roubik 1995). The pollination success of insect‐pollinated wild and cultivated plants is usually not dependent on a single, specialized pollinator species, but rather on a functionally diverse community of pollinators (Fontaine et al. 2006). Native pollinators, especially wild bees and hoverflies became increasingly important for crop production due to a continuing strong rise in the fraction of agriculture that depends on animal pollination (Hoehn et al. 2008; Jauker & Wolters 2008; Aizen and Harder 2009; Meyer et al. 2009). However, there is clear evidence in some areas of recent declines in wild as well as in domesticated pollinators, which can have important negative ecological and economic impacts (Potts et al. 2010). According to the analysis of Gallai et al. (2009), the decrease of insect pollinators will especially affect the production of fruits, vegetables and stimulants, representing important cash crops for small scale farmers especially in the tropics. The major threats for pollinator diversity and pollination services are the destruction and fragmentation of natural and semi‐natural habitats, and the intensification of agricultural landscapes (Kremen et al. 2002; Tscharntke et al. 2005; Steffan‐Dewenter & Westphal 2008). Pollinator richness and visitation rate on crops show general and significant exponential declines with increasing distance from natural habitats, more steeply in tropical than in temperate regions (Ricketts et al. 2008). In the tropics, natural forests are the key source habitats of species diversity, and it is confirmed that native bees from forest habitats promote pollination of coffee in nearby plantations (Ricketts 2004; Klein 2009). Knowledge on the relationships between insect pollinators and landscape structure, land use change and habitat quality in tropical regions is still limited. Most researches are restricted to bees, and other pollinator guilds have rarely been considered (Klein et al. 2003; Brosi et al. 2007; Hoehn et al. 2010). Therefore we conducted a study to 45
Chapter 3 Hoverflies & Wild bees investigate the diversity and distribution of two pollinator guilds, wild bees (Apidae) and hoverflies (Syrphidae), in the major land use types of a tropical landscape in southern Yunnan, China. The region is part of the ‘Indo‐Burma hotspot’, one of the 34 global hotspots exceptionally rich in biodiversity (Biodiversity Hotspots 2007). The study area represented a tributary valley of the Mekong River. Traditional land use systems are irrigated rice fields along the river courses and shifting cultivation systems on the slopes, but the largest proportion of the land area was covered with primary and secondary forest until about 30 years ago. Since then, large areas of forest have been successively transformed into commercial rubber monoculture plantations (Hevea brasiliensis). This land use change is representative of the development of tropical southern Yunnan. Detailed data from a typical subregion showed that between 1988 and 2006, rubber plantations increased from 12% of the total land cover to 46%, whereas forested areas dropped from 49% to 28% (Hu et al. 2008). Expansion of rubber plantations decreased mostly the area of tropical rainforests (Li et al. 2007). The objective of this study was to analyze species richness and composition of the hoverfly and wild bee communities in the predominant habitat types of the investigated landscape. These included natural forest plots, open land and agricultural land as well as rubber plantations of different ages, in order to assess the responses of the pollinator guilds to the recent changes in land use. The influence of flowering forb species on the studied insect groups was studied in details. 3.2 Materials and Methods 3.2.1 Study area and sampling localities The study was carried out in the Naban River valley (ca. 11 000 ha) within the Naban River Watershed National Nature Reserve (NRWNNR) in Xishuangbanna, southern Yunnan province, south‐west China (22° 10’ N and 100° 38’ E). The region represents the most northern part of the humid tropics in Asia with a climate influenced by Monsoon and three distinct seasons: cool‐dry (October‐January, with the lowest monthly temperature of 15°C in December), hot‐dry (February‐April, with the highest monthly temperature of 25 °C in April) and a rainy season (May‐September) with annual precipitation of almost 1600 mm. The natural vegetation of the study region is tropical rainforest, consisting of different types of evergreen and seasonal forest 46 Chapter 3 Hoverflies & Wild bees depends on topography and elevation (Cao et al. 2006; Lü et al. 2010). Secondary and primary forest plots and fragments are widespread in the study area, but most cultivated land is covered by rubber plantations. Valley bottoms are covered by rice fields, there are various fruit and vegetable crops around the small villages, and grassland fallows along the slopes. To represent the most typical habitat types of this landscape, we selected 13 sampling localities including forests, rubber plantations, fallows and open lands (See Fig. 2.1). Further descriptions are given in Table 3.1. 3.2.2 Field records Insect sampling was carried out using one Malaise trap (3.5 x 2.0 x 1.5m, length x width x height) (Townes 1962) with two collection ports facing east and west respectively at each of the 13 sites. The collecting bottles of the Malaise traps were filled two thirds with a mixture of anti‐freeze liquid of blue color containing glycol and ethanol. All traps were arranged centered at each of the 13 sampling sites. Table 3.1 Overview and description of the 13 sampling localities in the study area of the Naban River valley. Location Site code Site descriptions Mandian MD‐FO Secondary forest, closed canopy at 35 m Naban NB‐FO Secondary forest, closed canopy at 35 m Anmaxinzhai AM‐FO Secondary forest, closed canopy at 35 m Guomenshan GMS‐FO Primary forest, closed canopy at 35 m Forest Rubber plantations Mandian MD‐RU 5 years, trees 7 m high, open canopy Naban NB‐RU 8 years, trees 12 m high, open canopy Anmaxinzhai AM‐RU 20 years, trees 20 m high, closed canopy Shiyidui SYD‐RU 40 years, trees 30 m high, closed canopy Open land Naban NB‐OP Forest clearfell between NB‐FO and NB‐RU Anmaxinzhai AM‐OP Grassland on a ridge Guomenshan GMS‐OP Shrubland succession Fallow MD‐FA Rice field fallow Guomenshan GMS‐FA Rice field fallow Mandian 47
Chapter 3 Hoverflies & Wild bees Trap collections were conducted in different seasonal periods, including: (a) the beginning of the rainy season (May‐July 2008), (b) the beginning of the cool‐dry season (September‐November 2008), and (c) the transition period from the hot‐dry to the rainy season (March‐June 2009). At all sites, traps were emptied every 10 days of the collecting periods (with few exceptions where traps were destroyed or collection was impossible due to heavy rains). Insects were preserved in 70% ethanol for further identification to species level. Data analyses were based on the total number of morphological species and individuals from all sampling dates per sampling site. Vascular plant species from each of the 13 trap sites were recorded in March 2009. At the four natural forest sites, four 20 x 20 m2 plots per site were established around the trap locations to record the numbers of tree and liana species. The other plant species (representing the groundcover vegetation < 2 m) in the forest plots were recorded from four 5 x 5 m2 subplots within each of the four large plots. Total plant species richness of the four small and the four large plots were used for further calculations. In the other sites, records from all four 5 x 5 m2 plots per site provided the plant species numbers and abundances used for calculations. 3.2.3 Data analysis Cluster analyses were performed to identify quantitatively similar groups of hoverfly and wild bee communities among the different localities and habitats. One‐way ANOSIM (analysis of similarities) global tests were then applied to test for differences between groups created by the cluster analysis (Site code MD‐FO did not collect any hoverfly species, and it was excluded from this analysis). The tests were based on the Morisita index of similarity (Morisita 1959), which is recommended as the best overall similarity measure for abundance data by Wolda (1981) and Krebs (1998). Differences in total abundances and species number of hoverflies and wild bees between the groups produced by cluster analysis were compared by the Mann‐Whitney U‐test using Minitab 15.0 software (Minitab Inc., State College PA, USA). This non‐parametric test was applied because most of the data were found to be non‐normally distributed. Non‐metric multidimensional scaling (NMS; Kruskal 1964) using the Bray–Curtis index for abundance data after general relativization was applied to display and test for differences in hoverflies and wild bees community composition across the habitat 48 Chapter 3 Hoverflies & Wild bees types. This analysis was performed with PC‐ORD software (McCune & Mefford 2006). Eleven vegetation and land use variables were log transformed and included in the NMS analysis to test their effects on the hoverflies and bees (See Chapter2, Table 2.1). Six variables referred to plant species richness, including the total number of vascular plant species per site and the species numbers of different life forms, i.e. grasses, forbs (non‐woody dicotyledonous species), trees and lianas. The remaining variables were the degree of tree canopy cover and the degree of ground vegetation cover, the maximum vegetation height, the successional age of the study site (years after establishment of the present vegetation or age of the trees), and land use type with four categories, i.e. rice field fallows, open land (forest clear fell, grassland, scanty shrubland), rubber plantation and natural forest (Table 3.1). Correlation between the ordination and the environmental variables was calculated with the Pearson coefficient. Indicator Species Analysis based on the combined values of relative abundance and relative frequency of species (McCune & Grace 2002) was used to identify hoverflies and wild bees affiliated with specific land use types. The indicator value of each of the recorded species was calculated with PC‐ORD software (McCune & Mefford 2006) using 4999 runs in a Monte Carlo test considering values at P<0.05. The Chao 1 estimator was used to estimate total species richness of hoverflies and wild bees per site. This estimator is the sum of the observed number of species and the quotient a2/2b, where a and b equal the number of species represented by one (singletons) and two (doubletons) individuals, respectively. Calculations were conducted using the software package EstimateS (Version 8.0; Colwell 2006). To examine the relationship between Chao 1 estimated species richness of hoverflies and wild bees and the total of 11 vegetation and land use variables, these were included in the NMS analysis, Spearman’s Rho non‐parametric correlation was used. 3.3 Results Hoverflies A total of 1133 hoverfly individuals of 53 morphological species were recorded (Appendix 3.1). The cluster analysis dendrogram based on the quantitative similarities of the hoverfly communities was shown in Fig. 3.1. As supported by global one‐way ANOSIM tests, meaningful differences between communities occurred at a similarity of 0.6 and generated 3 subgroups (global R = 0.89; P < 0.001). Subgroup 1 was represented by five sites, with the three forest sites and the two oldest rubber 49
Chapter 3 Hoverflies & Wild bees plantations (20 and 40 years; site code MD‐FO did not collected any hoverfly species). Subgroup 2 included the two youngest rubber plantations (5 and 8 years) and three sites of open land, and subgroup 3 was formed by the two rice field fallows. Subgroup 1 had the lowest, and subgroup 3 had the highest mean value of both species and 1
N
BFO
AM
-F
S O
YD
-R
G
U
M
SA FO
M
-R
N U
BO
P
N
BR
A U
M
-O
G P
M
SM OP
D
-R
G U
M
SM FA
D
-F
A
0.8
Similarity (Morisita-Index)
1
0.6
2
3
0.4
0.2
Fig. 3.1 Quantitative similarity cluster analysis of the hoverfly communities at the sampling localities (site codes see Table 3.1), generated from the Morisita index using UPGMA. The dendrogram shows three subgroups at similarity levels of >0.6, indicated by the dashed line. (The forest site MD‐FO provided no hoverfly records and was excluded from the analysis). 20
70
a
b
15
c
10
b
5
a
0
Number of hoverfly individuals
Number of hoverfly species
60
50
40
c
30
20
b
10
a
0
1
2
3
Subgroup produced by cluster analysis
1
2
3
Subgroup produced by cluster analysis
Fig. 3.3 Number of (a) species and (b) individuals of hoverflies in the three subgroups produced by cluster analysis (1‐3, see Fig. 3.1). Box and whisker plots illustrate the 5th, 25th, 50th (median), 75th, and 95th percentiles, and the means as the dashed line. Different letters indicate significant differences (P < 0.05; Mann‐Whitney U‐Test). Small circle means the outlier. 50 Chapter 3 Hoverflies & Wild bees individuals (Fig. 3.3). Subgroup 2 shows intermediate values. The mean number of species and individuals was significantly different in all comparisons (P < 0.01; Mann‐Whitney U‐Test). Fig. 3.5a NMS ordination of hoverfly assemblages at the 13 sample sites: (▲) rubber plantation; (▽) grassland and shrubland; (◆) rice field fallow; (●) forest. Variables with a Pearson correlation coefficient at P<0.05 are shown. Cumulative variation in the original dataset explained by ordination is 77% (Axis 1 = 49.5%, Axis 2 = 27.5%, Final stress = 9.62, Final instability = 0.00001). Two axes were recommended by the NMS ordination of hoverfly communities (Fig. 3.5a) which together explained 77% of the total variance in the data set (Axis 1 = 49.5%, Axis 2 =27.5%) at a final stress of 13.5. The first axis strongly correlated with the number of forb species, while the second axis correlated with the vegetation height, canopy coverage, successional age and the number of liana species. Most of the species concentrated on the first quadrant and reached the highest scores on the first axis. Thus mostly of the species occurred at sites with a rich forbs species distributed (Fig. 3.5a). A total of five hoverfly species with indicator values at P < 0.05 were identified (see Appendix 3.1 for species indication, labeled with ‘*’). Of these, three species were specialists of rice field fallows and two species were specialists of 51
Chapter 3 Hoverflies & Wild bees open land. No hoverfly species were significantly affiliated with rubber plantations and forest. There was a highly positive relationship between the estimated species richness of hoverflies and the number of forb species at the sampling localities (r = 0.78; P < 0.01; Table 3.2). The other environmental variables including number of liana species, canopy coverage, vegetation height and successional age of sampling sites had a negative relationship with the estimated species richness of hoverflies. Table 3.2 r value of relationship between estimated Chao 1 species richness of hoverfly and wild bees and environmental variables within four major land‐use types of landscape in NRWNNR (n=13). Spearman’s Rho correlations are shown and the direction of the relationship indicated with + or – respectively. GroCov, CanCov, SucAge and VegHei means percentage of ground vegetation cover, percentage of canopy cover, the successional age of the site or age of the trees and vegetation height respectively. Group Plant Grass Forb Liana Shrub Tree GroCov CanCov VegHei SucAge
Hoverflies ‐0.14 0.46 0.78** ‐0.62* ‐0.52 ‐0.50
Wild bees 0.29 0.44 0.25 0.11
0.16
0.04
Spearman’s Rho correlation significance: *P<0.05; **P<0.01. 0.54 0.52 ‐0.59* ‐0.76** ‐0.61*
‐0.31 ‐0.34 ‐0.04
Wild bees A total of 1284 wild bee individuals of 44 morphological species were captured (Appendix 3.2). Meaningful differences between communities occurred at a similarity > 0.60 and generated two subgroups (global R = 0.74; P < 0.001) which was supported by global one‐way ANOSIM tests (Fig. 3.2). Subgroup 1 was formed by the three open land sites, two rice field fallows and two rubber plantations at the age of 8 and 40 years respectively. Subgroup 2 included four forest sites and two rubber plantations at the age of 5 and 20 years. However, species richness and abundance of wild bees did not differ between the two subgroups (Fig. 3.4, P =0.53; Mann‐Whitney U‐Test), species richness in the forest was significantly higher than in the other habitats. Two axes were also recommended by the NMS ordination of wild bee communities (Fig. 3.5b) which together explained 74.9% of the total variance in the data set (Axis 1 = 46.9%, Axis 2 = 28%) at a final stress of 12.9. The second axis strongly correlated only with the number of grass species. From the site and species points of NMS ordination diagram (Fig. 3.5b), most of the species sparsely distributed on the first, second and fourth quadrant. All environmental variables except the land use category showed significant effects at P<0.05 (Pearson correlation coefficient) for wild bees (Fig. 3.5b). Three wild bee species affiliated with forest species with indicator values at P < 0.05 were identified, and no meaningful indicator species affiliated with the 52 Chapter 3 Hoverflies & Wild bees other three land use types were found (see Appendix 3.2 for species indication, labeled with ‘*’). No significant relationship between the estimated species richness of wild bees and 11 environmental variables was found (P > 0.05; Table 3.2). 1
G
M
S
SY -OP
D
G RU
M
S
AM - F A
-O
N P
BR
N U
BO
M P
D
-F
AM A
-F
M O
D
-R
M U
D
-F
G O
M
S
N - FO
BF
AM O
-R
U
0.8
Similarity (Morisita-Index)
1
2
0.6
0.4
0.2
Fig. 3.2 Quantitative similarity cluster analysis of the wild bee communities at the sampling localities (site codes see Table 3.1), generated from the Morisita‐index using UPGMA. The dendrogram shows two subgroups at similarity levels of >0.60, indicated by the dashed line. 70
20 a
b
50
40
10
Number of wild bee individuals
Number of wild bee species
60
15
a
30
a
5
20
a
0
0
a
10
1
2
Subgroup produced by cluster analysis
1
2
Subgroup produced by cluster analysis
Fig. 3.4 Number of (a) species and (b) individuals of wild bees in the two subgroups produced by cluster analysis (1 and 2, see Fig. 3.2). Box and whisker plots illustrate the 5th, 25th, 50th (median), 75th, and 95th percentiles, and the means as the dashed line. Different letters indicate significant differences (P < 0.05; Mann‐Whitney U‐Test). Small circle means the outliners. 53
Chapter 3 Hoverflies & Wild bees Fig. 3.5b NMS ordination of wild bee assemblages at the 13 sample sites: (▲) rubber plantation; (▽) grassland and shrubland; (◆) rice field fallow; (●) forest. Variables with a Pearson correlation coefficient at P<0.05 are shown. Cumulative variation in the original dataset explained by ordination is 74.9% (Axis 1 = 46.9%, Axis 2 = 28%, Final stress = 12.9, Final instability = 0.00001). 3.4 Discussion The two guilds of pollinators showed contrasting patterns of species distribution and abundance in the surveyed landscape. Hoverflies Hoverflies were most common in young successional stages with highest number in the rice field fallows and rare in forests, and species richness was closely correlated with the number of flowering forb species in the different habitat types. This relationship can be explained by the resources provided by the plants, i.e., nectar and pollen serving as food for the adults of all hoverfly species. Nectar food resource is an important environmental variable for insect pollinators. The species richness 54 Chapter 3 Hoverflies & Wild bees and distribution of flowering plants are found to be the major factors in determining pollinator diversity in various landscapes of Europe (Batáry et al. 2010; Kleijn & van Langevelde 2006; Meyer et al. 2009; Steffan‐Dewenter & Tscharntke 2001). In our study area, hoverflies species were also showed such close relationship with food resources. Flowering forb species can be considered as indicators for the “openness” of a habitat, as they usually decreased with proceeding succession from agricultural land to forest as well as from young to old rubber plantations with increasing canopy coverage. This was confirmed by the result that species number and abundance of hoverflies are higher in young rubber plantations (5 and 8 years) than in older ones (20 and 40 years) and in forests. Similar patterns of hoverfly species richness and abundance are recorded from various other landscapes in temperate regions. The hoverflies species richness is not only concentrate on the most rewarding resources available in the landscape (Haenke et al. 2009), but also affected by other factors related to resource heterogeneity such as species richness of flowering plants, flower abundance, habitat area and landscape diversity (Kleijn & van Langevelde 2006; Meyer et al. 2009). Though study by Jauker et al. (2009) shows that species richness of hoverflies do not decline with increasing distance from the main habitat within an agricultural landscape, the majority (nearly 80%) of the hoverfly species recorded are associate with open space habitats rather than closed‐canopy forest (Gittings et al. 2006). This figure was very similar to the result from the present study area, where 41 (77%) of the 53 recorded species were found in the rice field fallows. Overall, the general conclusion from the studies in temperate regions that hoverfly species richness is usually highest in open land and especially in habitats with high supply of flowering resources, can also be drawn from the present study in a tropical landscape. The natural tropical forest was not the habitat source of the vast majority of the hoverflies recorded in this area. Rather, its reduction by land cultivation favored the richness and distribution of hoverflies, indicating that most of the species collected might not originate from forest sites. Wild bees In contrast to the hoverflies, wild bees showed significantly higher numbers of species and individuals in forests than in any other types of habitat. The latter had high similarities in bee community composition among each other, as indicated by the cluster analysis (subgroup 2 in Fig. 3.2), but no close correlation with flowering 55
Chapter 3 Hoverflies & Wild bees forbs and other environmental variables (Table 3.2). Although 30 (68%) of 44 species in total were recorded from at least one of the forest sites, most of these species were additionally recorded from other habitat types. This indicated that many bee species showed low specificities for habitat and floral resources and high abilities to move within the landscape. The relationships between floral resources and species diversity of bees are also not confirmed in other tropical areas. For example, no consistent differences in bee diversity or abundance with respect to pasture management or floral resources were found by Brosi et al. (2007) and Klein et al. (2003). However, Hoehn et al. (2010) found that bee density and diversity in a tropical landscape in Indonesia is highest in managed systems and concluded that the herb associated bee community profits from the opening of the landscape as a result of agricultural activities. Small patches of natural habitat dispersed in cultivated landscapes may support high bee abundance even in regions with low proportions of natural habitat (Winfree et al. 2008). Forest fragment size or disturbance are not important factors for determining bee diversity and abundance in tropical areas (Tonhasca et al. 2002; Brosi et al. 2008), and bee species richness is even higher in disturbed compared to primary forest in Southeast Asia (Liow et al. 2001). Strong forest disturbance, however, is found to reduce species richness of stingless bees (Cairns et al. 2005). In contrast to the results of the present study, Hoehn et al. (2010) recorded higher local bee density and diversity in open land than in primary forest in Indonesia, but highest overall bee richness in agroforestry systems is due to high crop diversity. However, direct comparison of these results is difficult, because they are based on different sampling methods, and are partly based on selected taxa of bees. Despite this, those studies indicate that wild bee diversity and distribution patterns in tropical landscapes are complex and often affected by combinations of different factors. Although floral resources are of general importance, their relationship to bees can be influenced by landscape patterns and disturbance effects. In addition, wild bee communities are composed of species with different habitat and resource requirements, due to differences in the use of flowering plants, nesting requirements, dispersal modes, and other traits of species related to landscape structure (Steffan‐Dewenter & Tscharntke 2001; Steffan‐Dewenter et al. 2002; Potts et al. 2003; Jauker et al. 2009). 56 Chapter 3 Hoverflies & Wild bees Implications of land use change by rubber cultivation The present land use types of the study area showed a pattern of closed and open habitats with dynamic spatial and temporal changes due to rubber cultivation. Currently, most of the rubber plantations are less than 20 years old, with mainly open canopies and a relatively rich diversity of flowering herbs. At these stages, the plantations provided suitable habitats for large subgroups of the hoverfly guild (subgroup 2 in Fig. 3.1) as well as for most of the bees recorded from open land (subgroup 1 in Fig. 3.2). With increasing plantation age of about 40 years before felling, habitat quality decreased due to the successive reduction of the ground cover vegetation under the closing canopy. A further expansion of rubber cultivation will result large areas of mature rubber plantations. Because rubber cultivation largely proceeded at the expense of forest areas and not agricultural land, we assume that hoverfly communities will not be negatively affected by this development. However, about a quarter of wild bee species were only recorded from forests, indicating that natural forest habitats were necessary to sustain the populations of many wild bee species. The large number of wild bee species which were recorded from forest can be negatively affected in a landscape with increasing rubber plantations. Nonetheless, effects of rubber cultivation on hoverflies can be mitigated by a management of the rotation cycles that allows the steady presence of young plantation stages with ground cover vegetation. It was that can serve as a temporal habitat providing flowering resources. Though larvae of hoverflies can eat decaying plant and animal matter to feeding in aphids and thrips, the wild bees are central place foragers and their larval development depends on the honey. The composition of wild bee communities is generally more heterogeneous in terms of the ecological requirements compared to hover flies. Therefore more detailed analyses of the life history traits of different bee taxa would be necessary to identify the factors shaping their distribution patterns. This study confirmed that the wild bees and hoverflies had the contrasting responses to a dynamic changing tropical area of southern Yunnan. 3.5 References Aizen A, Harder LD (2009). The global stock of domesticated honey bees is growing slower than agricultural demand for pollination. Current Biology, 19, 915‐918. Batáry P, Báldi A, Sáropataki M, Kohler F, Verhulst J, Knop E, Herzo F, Kleijn D (2010). Effect of conservation management on bees and insect‐pollinated grassland plant 57
Chapter 3 Hoverflies & Wild bees communities in three European countries. Agriculture, Ecosystems and Environment, 136, 35‐39. Biodiversity Hotspots (2007) <http://www.biodiversityhotspots.org/> Brosi BJ, Daily GC, Ehrlich PR (2007). Bee community shifts with landscape context in a tropical countryside. Ecological Applications, 17, 418‐430. Brosi BJ, Daily GC, Shih TM, Oviedo F, Duran G (2008). The effects of forest fragmentation on bee communities in tropical countryside. Journal of Applied Ecology, 45, 773‐783. Cairns CE, Villanueva‐Gutierrez R, Koptur S, Bray DB (2005). Bee populations, forest disturbance, and africanization in Mexico. Biotropica, 37, 686‐692. Cao M, Zou X, Warren M, Zhu H (2006). Tropical forests of Xishuangbanna, China. Biotropica, 38, 306‐309. Colwell RK (2006). EstimateS: Statistical estimation of species richness and shared species from samples. Version 8.0. Department of Ecology and Evolutionary Biology, University of Connecticut, Storrs, USA, <http://viceroy.eeb.uconn.edu/estimates>. Fontaine C, Dajoz I, Meriguet J, Loreau M (2006). Functional diversity of plant‐pollinator interaction webs enhances the persistence of plant communities. PLoS Biology, 4, 129‐135. Gallai N, Salles JM, Settele J, Vaissiere BE (2009). Economic valuation of the vulnerability of world agriculture confronted with pollinator decline. Ecological Economics, 68, 810‐821. Gittings T, O'Halloran J, Kelly T, Giller PS (2006). The contribution of open spaces to the maintenance of hoverfly (Diptera, Syrphidae) biodiversity in Irish plantation forests. Forest Ecology and Management, 237, 290‐300. Haenke S, Scheid B, Schaefer M, Tscharntke T, Thies C (2009). Increasing syrphid fly diversity and density in sown flower strips within simple vs. complex landscapes. Journal of Applied Ecology, 46, 1106‐1114. Hoehn P, Steffan‐Dewenter I, Tschantke T (2010). Relative contribution of agroforestry, rainforest and openland to local and regional bee diversity. Biodiversity and Conservation, 19, 2189‐2200. Hoehn P, Tschantke T, Tylianakis JM, Steffan‐Dewenter I (2008). Functional group diversity of bee pollinators increases crop yield. Proceedings of the Royal Society B, Biological Sciences, 275, 2283‐2291. Hu HB, Liu WJ, Cao M (2008) Impact of land use and land cover changes on ecosystem services in Menglun, Xishuangbanna, Southwest China. Environmental Monitoring and Assessment, 146, 147‐156. Jauker F, Diekotter T, Schwarzbach F, Wolters V (2009). Pollinator dispersal in an agricultural matrix: opposing responses of wild bees and hoverflies to landscape structure and distance from main habitat. Landscape Ecology, 24, 547‐555. Jauker F, Wolters V (2008). Hover flies are efficient pollinators of oilseed rape. Oecologia, 156, 819‐823. Kleijn D, van Langevelde F (2006). Interacting effects of landscape context and habitat quality on flower visiting insects in agricultural landscapes. Basic and Applied Ecology, 7, 201‐214. Klein AM (2009). Nearby rainforest promotes coffee pollination by increasing 58 Chapter 3 Hoverflies & Wild bees spatio‐temporal stability in bee species richness. Forest Ecology and Management, 258, 1838‐1845. Klein AM, Vaissiere BE, Cane JH, Steffan‐Dewenter I, Cunningham SA, Kremen C, Tscharntke T (2007). Importance of pollinators in changing landscapes for world crops. Proceedings of the Royal Society B, Biological Sciences, 274, 303‐313. Klein AM, Steffan‐Dewenter I, Tscharntke T (2003). Fruit set of highland coffee increases with the diversity of pollinating bees. Proceedings of the Royal Society B, Biological Sciences, 270, 955‐961. Krebs Ch (1998). Ecological Methodology 2nd Edition. Benjamin/Cummings, Menlo Park, CA. Kremen C, Williams NM, Thorp RW (2002). Crop pollination from native bees at risk from agricultural intensification. Proceedings of the National Academy of Sciences of the United States of America, 99, 16812‐16816. Kruskal JB (1964). Nonmetric multidimensional scaling: a numerical method. Psychometrika, 29, 115‐129. Li H, Aide TM, Ma Y, Liu W, Cao M (2007). Demand for rubber is causing the loss of high diversity rainforest in SW China. Biodiversity and Conservation, 16, 1731‐1745. Liow LH, Sodhi NS, Elmqvist T (2001). Bee diversity along a disturbance gradient in tropical lowland forests of south‐east Asia. Journal of Applied Ecology, 38, 180‐192. Lü XT, Yin JX, Tang JW (2010). Structure, tree species diversity and composition of tropical seasonal rainforests in Xishuangbanna, south‐west China. Journal of Tropical Forest Science, 22, 260‐270. McCune B, Grace J (2002). Analysis of ecological communities. MjM Publishers, Gleneden Beach, Oregon, USA. McCune B, Mefford MJ (2006). PC‐ORD. Multivariate analysis of ecological data. Version 5.10. MjM Software, Gleneden Beach, Oregon, USA. Meyer B, Jauker F, Steffan‐Dewenter I (2009). Contrasting resource‐dependent responses of hoverfly richness and density to landscape structure. Basic and Applied Ecology, 10, 178‐186. Morisita M (1959). Measuring of interspecific association and similarity between communities. Mere. Fac. Sci. Kyushu Univ. Set. E (Biol.), 3, 65‐80. Potts SG, Biesmeijer JC, Kremen C, Neumann P, Schweiger O, Kunin WE (2010). Global pollinator declines: trends, impacts and drivers. Trends in Ecology and Evolution, 25, 345‐353. Potts SG, Vulliamy B, Dafni A, Ne'eman G, Willmer P (2003). Linking bees and flowers: How do floral communities structure pollinator communities? Ecology, 84, 2628‐2642. Ricketts TH (2004) Tropical forest fragments enhance pollinator activity in nearby coffee crops. Conservation Biology, 18, 1262‐1271. Ricketts TH, Regetz J, Steffan‐Dewenter I, Cunningham SA, Kremen C, Bogdanski A, Gemmill‐Herren B, Greenleaf SS, Klein A‐M, Mayfield MM, Morandin LA, Ochieng A, Viana BF (2008). Landscape effects on crop pollination services: are there general patterns? Ecology Letters, 11, 499‐515. Roubik DW (1995). Pollination of cultivated plants in the tropics. FAO Agricultural Services Bulletin, 118. Steffan‐Dewenter I, Munzenberg U, Burger C, Thies C, Tscharntke T (2002). Scale‐dependent 59
Chapter 3 Hoverflies & Wild bees effects of landscape context on three pollinator guilds. Ecology, 83, 1421‐1432. Steffan‐Dewenter I, Tscharntke T (2001). Succession of bee communities on fallows. Ecography, 24, 83‐93. Steffan‐Dewenter I, Westphal C (2008). The interplay of pollinator diversity, pollination services and landscape change. Journal of Applied Ecology, 45, 737‐741. Tonhasca A, Blackmer JL, Albuquerque GS (2002). Abundance and diversity of euglossine bees in the fragmented landscape of the Brazilian Atlantic forest. Biotropica, 34, 416‐422. Townes H (1962). Design for a Malaise trap. Proceedings of the Entomological Society of Washington, 64, 253‐262. Tscharntke T, Klein AM, Kruess A, Steffan‐Dewenter I, Thies C (2005). Landscape perspectives on agricultural intensification and biodiversity ‐ ecosystem service management. Ecology Letters, 8, 857‐874. Winfree R, Williams NM, Gaines H, Ascher JS, Kremen C (2008). Wild bee pollinators provide the majority of crop visitation across land‐use gradients in New Jersey and Pennsylvania, USA. Journal of Applied Ecology, 45, 793‐802. Wolda H (1981). Similarity indices, sample size and diversity. Oecologia, 50, 296‐302. 60 Chapter 3 Hoverflies & Wild bees Appendix 3.1 List of hoverfly species and numbers of individuals recorded from the 13 study localities compiled by habitat types (* means indicator species). Species Forest Rubber
Rubber Open Rice (5‐8 y.) (20‐40 y.) land fallow Asarkina cf. porcina Coquillett 1898 Baccha sp. Chrysotoxum sp. Citrogramma citrinum Brunetti 1923 Dideopsis aegrota Fabricius 1805 Episyrphus alternans Macquart 1842 Episyrphus sp. Eristalodes paria Bigot 1880 Eumerus sp. 1* Eumerus sp. 2 Eumerus sp. 3 Eumerus sp. 4* Helophilus affinis Wahlberg 1844 Lathyophthalmus quinquestriatus Fabricius 1794
Lathyophthalmus sp. 1 Melanostoma sp. 1 Microdon sp. 1 Microdon sp. 2 Microdon sp. 3 Milesiinae Gen. sp. 1 Milesiinae Gen. sp. 2 Milesiinae Gen. sp. 3 Milesiinae Gen. sp. 4 Milesiinae Gen. sp. 5 Milesiinae Gen. sp. 6 Paragus spec. 1 Phytomia errans Fabricius 1787 Platycheirus sp. 1 Platycheirus sp. 2 Platycheirus sp. 3* Platycheirus sp. 4 Platycheirus sp. 5* Sphaerophoria sp. 1 Sphaerophoria sp. 2* Sphaerophoria sp. 3 Sphaerophoria sp. 4 Sphaerophoria sp. 5 Sphaerophoria sp. 6 Syritta sp. 1 0 0 0 0 1 0 2 0 1 0 0 0 0 0 0 0 0 0 0 56 7 3 3 19 8 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4 10 24 0 3 0 0 0 1 0 0 0 1 0 0 23 19 0 0 20 3 0 0 0 0 0 0 0 0 0 0 1 0 1 0 0 1 0 0 5 1 1 0 0 0 0 0 0 0 0 0 0 0 0 23 4 0 0 12 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 2 0 4 12 21 0 23 1 1 24 0 2 0 2 0 1 0 55 47 0 0 63 31 3 0 11 0 2 2 0 11 5 7 4 3 5 0 2 0 5 2 3 6 9 2 2 1 1 2 0 1 1 14 0 0 1 5 27 0 0 37 6 5 1 41 1 50 5 2 44 70 6 36 27 10 2 61
Chapter 3 Hoverflies & Wild bees Syritta sp. 2 Syrphinae Gen. sp. Syrphus sp. 1 Syrphus sp. 2 Syrphus sp. 3 Syrphus sp. 4 Syrphus sp. 5 Syrphus sp. 6 Syrphus sp. 7 Syrphus sp. 8 Xylota cf. coquilletti Hervé‐Bazin 1914 Xylota sp. 1 Xylota sp. 2 Xylotini Gen. sp. 0 1 0 0 1 0 1 0 0 0 0 2 0 14 0 0 0 0 0 1 0 0 0 0 0 0 0 6 0 0 0 0 0 0 0 0 0 0 0 0 0 5 0 0 0 3 1 2 10 1 2 1 2 0 1 30 1 0 2 4 4 2 0 0 2 2 0 0 0 2 Appendix 3.2 List of wild bee species and numbers of individuals recorded from the 13 study localities compiled by habitat types (* means indicator species). Species Forest
Rubber
Rubber
Open Rice (5‐8 y.) (20‐40 y.)
land fallow Amegilla calceifera Cockerell 1911 Andrena sp. 1* Anthophora sp. 1 Anthophora sp. 2 Anthophora sp. 3 Anthophora sp. 4 Apis andreniformis Smith 1858* Apis cerana Fabricius 1793 Apis dorsata Fabricius 1793 Bombus friseanus Skorikov 1933 Bombus impetuosus Smith 1871 Ceratina cognate Smith 1879 Ceratina flavipes Smith 1879 Ceratina laeviuscula Wu 1963 Coeliaxys sp. 1 Coeliaxys sp. 2 Cyaneoderes caerulea Fabricius 1804 Cyaneoderes tumida Friese 1903 Dianthidium chinensis Wu 1962 Halictus grandiceps Cameron 1897 Lasioglossum albescens Smith 1853 Lasioglossum subopacum Smith 1853 Lipotriches sp. 1 62 42 3 2 7 24 1 10 18 4 40 0 13 5 18 26 4 8 3 7 2 57 2 3 22 0 0 1 30 1 0 22 2 21 0 4 6 1 7 7 1 1 0 5 35 0 7 4 0 0 3 5 0 0 2 0 4 0 0 1 1 6 0 0 2 0 1 13 0 2 5 0 0 0 2 0 0 27 0 8 0 22 5 15 15 3 0 0 1 6 77 2 13 2 0 0 0 0 0 0 35 0 11 1 15 4 7 5 1 3 3 0 13 42 1 5 Chapter 3 Hoverflies & Wild bees Lipotriches sp. 2 Lipotriches sp. 3 Lipotriches sp. 4 Lipotriches sp. 5 Macropis sp. 1 Megachile conjunctiformis Yasumatsu 1938
Megachile monticola Smith 1853 Megachile velutina Smith 1853 Nomada sp. 1 Nomia ellioti Smith 1875 Nomia fuscipennis Smith 1875 Nomia iridescens Smith 1853* Nomia punctulata Dalla Torre 1896 Nomia yunnanensis Wu 1983 Nomia sp. 1 Nomia sp. 2 Stelis sp. 1 Thyreus sp. 1 Trigona iridipennis Smith 1854 Xylocopa auripennis Lepeletier 1841 Xylocopa nasalis Westwood 1842 10 0 26 0 1 0 0 1 1 9 12 5 0 5 3 0 0 1 0 2 2 24 2 31 0 0 0 0 0 1 5 3 0 0 8 0 0 0 1 0 1 0 6 0 4 0 0 0 1 1 0 0 1 0 1 3 0 0 0 0 0 0 2 67 2 53 5 0 1 3 9 0 0 8 0 1 15 4 1 1 3 3 0 1 20 1 25 2 0 0 6 5 0 0 2 0 0 3 0 1 1 0 3 0 0 63
Chapter 3 Hoverflies & Wild bees 64 Chapter 4 Seed dispersal of Musa acuminata Chapter 4 Spatial and temporal effects on seed dispersal and seed predation of Musa acuminata in southern Yunnan, China
Meng, L.‐Z., Gao, X.‐X., Chen, J., K. Martin, (2012). Spatial and temporal effects on seed dispersal and seed predation of Musa acuminata in southern Yunnan, China. Integrative Zoology, DOI: 10.1111/j.1749‐4877.2011.00275.x. 65
Chapter 4 Seed dispersal of Musa acuminata Abstract Why is wild banana (Musa acuminata) very abundant and why is the succession process quicker following increasing disturbance in the tropical area has been noticed by many ecologists. This study was conducted to analyze animal‐seed interactions and their effects on the seed fate of a wild banana species in tropical southern Yunnan (China) through experiments considering spatial (site and habitat) and temporal (seasons) variation. Only 13% of the fruits were removed by climbing seed predators (different species of rats). The largest proportion of fruits (81%) was removed by frugivorous seed dispersers, especially by bats at night‐time. In the ant exclosure treatment, 69% of all artificially exposed seeds were removed, compared to 56% of all seeds in the rodent exclosure treatment. That is, rodents accounted for a significantly higher total seed removal rate than ants, but with spatial and temporal differences. The highest seed predation rate by rodents (70%) was found in forest with wild banana stands, corresponding with the highest rodent diversity (species numbers and abundance) among the habitat types. In contrast, the seed removal rate by ants (57%) was highest in the open land habitats, but there was no close correlation with ant diversity. Seed removal rates by ants were significantly higher in the dry compared to the rainy season, but rodent activity showed no differences between seasons. The overall results suggest that the largest proportion of seeds produced by wild banana are primarily dispersed by bats. Primary seed dispersal mostly by bats at night time is essential for wild banana seeds to escape seed predation. Key words: Bats, Musa acuminata, seed dispersal, seed predation, tropical forest. 66 Chapter 4 Seed dispersal of Musa acuminata 4.1 Introduction Seed dispersal is the process by which plant seeds and other plant propagules move from the parent plants to settle in a more or less distant area (Herrera 2002). About 80% of the woody plant species of tropical forests produce fleshy fruits to attract animals (frugivores) as seed dispersers (Jordano 1992). Dispersal distances by animals range from a few meters (e.g. by ants) to more than ten kilometers (e.g. by some birds, bats or large mammals) (Corlett 2009). Different hypotheses have been used to explain the principal benefits of seed dispersal to plants, which include (1) the escape from seed‐dependent or distance‐responsive seed predators, pathogens or seedling competition near the parent plant (Janzen 1970; Connell 1971); (2) the colonization of a suitable habitat at a relatively large distance from the parent plants (Baker 1974), and (3) the directed dispersal via some non‐random process to specific sites that offer a disproportionately high probability of seedling establishment (Davidson & Morton 1981). Overall, the process of animal seed dispersal is complex and covers two broad categories or phases (Herrera 2002; Vander Wall & Longland 2004; Vander Wall et al. 2005a). The first category or phase (primary dispersal) includes the seed removal from the parent plant which can occur when a seed/fruit falls directly from the parent to the ground or, in the case of a seed is taken directly from a parent tree by a vertebrate frugivore. Seeds are then released by defecation and regurgitation in different densities (Medellín & Gaona 1999; Thomas et al. 1988). The second category or phase (secondary dispersal) includes the processes leading to the end‐point of the seed and may include dispersal mechanisms by animals on the ground (insects, rodents and birds), which remove seeds from feces or from the soil (Herrera 2002; Vander Wall et al. 2005a). Abiotic factors such as transport by water are also often involved in the second category of seed dispersal (Schupp et al. 2010). However primary dispersal can also occur when a seed is eaten by a ground dwelling frugivore e.g. cassowary, after it has simply fallen to the ground. After this initial frugivore dispersal, secondary dispersal can now occur. Seed predation is another important process which affects the final seed fate (Wang & Smith 2002). The proportion of seed predation is influenced by the combination of many factors such as different predators (Hulme 1994), season (Myster 2003), site (García‐Castaño et al. 2006), seed density (von Allmen et al. 2004), seed traits (Brewer 2001), and others 67
Chapter 4 Seed dispersal of Musa acuminata (Janzen 1971). Previous studies showed that secondary seed dispersal is a precondition for seeds to escape seed predation (Vander Wall & Longland 2004). For example, secondary seed dispersal by ants (Levey & Byrne 1993) and dung beetles (Shepherd & Chapman 1998) can effectively help the seed to escape the predation from rodents. However, little is known on the role of ants in seed dispersal in tropical Asia (Corlett 2009). In some cases, rodents are active in both seed predation and secondary seed dispersal by their scatter‐hoarding behavior (Forget & Milleron 1991). In total, the seed fate and finally the chance of successful seedling establishment depend on the various mechanisms of primary dispersal, secondary dispersal and seed predation. Although there are many studies dealing with seed dispersal and seed predation under various ecological aspects (see Schupp et al. 2010 and Kolb et al. 2007 for reviews), relatively few studies have been aimed at linking the effects of different processes and species, but see Sivy et al. (2011), Vander Wall et al. (2005a; b), Christianini and Oliveira (2010), Ruiz et al. (2010) and Rodríguez‐Pérez and Travest (2010). To contribute to further understanding of animal‐seed interactions, we conducted a study on the processes affecting the fate of seeds of a wild banana (Musa acuminata Colla 1820) in tropical southern Yunnan (China) under consideration of different spatial (site and habitat) and temporal (seasons) factors. Specific objectives were (1) to estimate the proportion of fruits of M. acuminata removed by ground and volant animals (Birds at day time and bats at night time) in the natural habitat, (2) to estimate the proportion of seeds removed by secondary seed dispersers and seed predators in different habitats and seasons, and (3) to identify the species and species diversity of seed dispersers and seed predators in the different study sites and habitats.
4.2 Material and Methods 4.2.1 Characteristics of Musa acuminata Musa acuminata (Musaceae) is native to Southeast Asia and is a common pioneer species in many tropical regions including southern China, India, Laos, Myanmar, Thailand and Vietnam. The plants reach a height of more than 4 m and occur in various habitats with adequate light and moisture conditions, ranging from open land to forest. New habitats are colonized by seeds and further occupied by reproduction via root rhizomes. The plants can form patchy stands that dominate the pioneer plant 68 Chapter 4 Seed dispersal of Musa acuminata community. In the study area of southern Yunnan, wild banana produces fruits the whole year round with a peak in May, the beginning of the rainy season. The infructescence is about 1 m long and the single fruit has a size of about 9 cm in length. Each single fruit produces 70‐80 seeds with a size of 5‐6 x 3 mm. The seeds possess a warty appendage referred to as a caruncle. It represents a specific kind of elaiosome which is attractive to and consumed by ants. (Flora of China [eFloras.org]; Shugart 1998; Liu 2001). 4.2.2 Study area and sampling sites Field studies were conducted at three sites in the Dai autonomous prefecture of Xishuangbanna, southern Yunnan province, south‐west China. The region represents the northernmost part of the humid tropics in Asia with a climate influenced by Monsoon and three distinct seasons: cool‐dry (October‐January, with the lowest monthly temperature of 15°C in December), hot‐dry (February‐April, with the highest monthly temperature of 25 °C in April) with low precipitation and humidity, and a rainy season (May‐September) with most of the mean annual precipitation of almost 1600 mm. The natural vegetation of the study region is tropical rain forest, falling into different types of evergreen and seasonal forests related to topography and elevation (Cao et al. 2006; Lü et al. 2010). This study was carried out in the three sites in Xishuanbanna and each site was apart a straight distance of 70‐100km from each other. Each site contains 3 types of habitats including 1) forest habitat with wild banana, 2) forest habitat without banana, and 3) open land without wild banana closing to the forest edge representing degraded areas sparsely occupied by shrubs with a maximum height of one meter. The three habitat types per site were located at distances between 300 and 500 m from each other. Site 1 is a forest area in a valley, located at about 570 m asl near the Xishuangbanna Tropical Botanical Garden (XTBG; 21°56′N, 101°15′E) with an area of 7‐8 km2. It represents a disturbed fragment of a tropical seasonal lowland rain forest with a maximum tree height of 40‐50 meters. In this forest plot, a population of wild banana covered an area of about 0.2‐0.3 km2 with approximately 2000‐3000 stems. Site 2 is a slope forest area, located at about 960 m asl at Xinshan (21°50′N, 101°33′E) with an area of 5‐6 km2. It represents a secondary tropical montane forest of about 30 years with a maximum tree height of 25 meters. In this forest plot, a population of 69
Chapter 4 Seed dispersal of Musa acuminata wild banana covered an area of about 0.3‐0.4 km2 with approximately 3,000‐4,000 stems. Site 3 (Bubeng, 21°37′N, 101°35′E) is a forest area in a valley, located at about 680 m asl with an area of 9‐10 km2. It represents an intact forest patch of a tropical seasonal lowland rain forest with an emergent tree height of 50‐70 meters (Lü et al.2010). In this forest plot, a population of wild banana covered an area of about 0.4‐0.5 km2 with approximately 4,000‐5,000 stems. 4.2.3 Experiments Experiment 1: Primary seed removal To determine the numbers of ripe fruits removed from wild banana plants by different animals the following treatments were conducted in the forest area of Site 1 by using 8 plants each: (a) Untreated plants with open access to all animals served as control; (b) to exclude climbing animals, a round iron sheet of 0.5 m in diameter was fixed around the stem of plants at 1 m start above the ground; (c) to exclude volant animals plants were individually fully enclosed to the top of plants by black nylon nets (mesh size less than 20 mm) supported by bamboo sticks to avoid contact with the plant. To allow climbing animals the access to the fruits, the net cover did not touch the ground. The numbers of removed fruits per treatment were counted in the morning at 07:00 hrs and in the evening at 19:00 hrs over a period of 30 days (April 15th to May 14th in 2004) at the end of dry season, and the number of fruits was also counted before they were removed. More than 75% single fruit that had been eaten was counted as “1 fruit removed”. Mean numbers of fruits per plant prior to the experiment were (42±1, means±SE; n=24). The fruits on all trees were developed at more or less same stage and equally attractive to animals. Additionally, three days continuous observation (April 20th to 22nd in 2004) from 07:00hrs to 19:00hrs on the activities of frugivorous birds was conducted with the untreated plants to record the primary seed dispersers during day time. Experiment 2: Secondary seed dispersal and predation To determine the numbers of wild banana seeds removed by secondary dispersers and seed predators, fresh seeds were experimentally exposed on the ground in split plot design with 3×3 for site in main plot and habitat in subplot. In all habitat types, 70 Chapter 4 Seed dispersal of Musa acuminata three treatments were conducted. In each treatment 10 fresh seeds each were placed on the flat surface of 10 inverted Petri dishes (12 cm in diameter) which were pressed in the soil flush to the ground. (a) To exclude ants, the fringes of the dish surface were smeared with grease. (b) To exclude rodents, the dish surface was protected by a wire mesh with a mesh size of less than 1 cm which allowed access by ants. (c) Seeds on the flat dish surface with open access to all dispersers and predators served as control. Totally 630 seeds per treatment, 210 seeds per site and 70 seeds per habitat were used. The dry and rainy season had 360 and 270 seeds artificially exposed respectively. The 10 dishes per treatment were located at a distance of 1 m from each other and were separated from the dishes of the other treatments at distances of 5 m. The experiments were conducted from March 14th to April 26th in the dry season and were repeated in the rainy season from September 5th to October 7th in 2004. Numbers of seeds in all dishes were counted at 08:00 hrs every day. The remaining seeds were exchanged by fresh seeds every three days. Experiment 3: Identification of seed dispersers To record rodent species and their abundance, 100 snap traps (length×wide×height: 23cm×12cm×10cm) baited with peanut per habitat were established in four rows of 25 traps each and distances of 5 m between traps and 25 m between rows. Traps were established in the field for 5 days (April 27th to May 1st in 2004) in the dry season and controlled at 07:00 and 19:00 hrs every day. The recording was repeated in the rainy season from October 8th to October 12th of 2004. Pitfall traps were used to record ants. Pitfall traps were plastic pots with a diameter of 7.5 cm and a depth of 13 cm which were buried flush to the soil surface, one third filled with 10% formalin solution. At each of the three habitat types per site, ten pitfall traps were arranged at a distance of ca. 5 m from each other. Traps were emptied every three days and totally three times in the dry season and in the rainy season each within the same study period of experiment 2. 4.2.4. Statistical analysis All data were analyzed using SPSS statistical software 13.0 version (SPSS Inc. 2004). The chi square test (χ2) was used to analyze the differences in seed removal rates with the different treatments in the experiment 1 and to determine what is the major primary seed disperser of wild banana at day and night time respectively. Paired t test was used to detect the difference of fruit removal rate between the treatment and 71
Chapter 4 Seed dispersal of Musa acuminata control. Multivariate analysis with Duncan’s multi‐comparisons method was used to analyze the effects of season, site and habitat in the secondary seed removal experiments. The t test was used to test for differences in treatments. 4.3 Results 4.3.1. Primary seed dispersal and predation The proportion of fruits removed from the wild banana plants at night were 75% in the control, 81% in the climbing animal exclusion and 13% in the volant animal exclusion treatments (Fig. 4.1). The proportion of fruits removed at night was significantly lower in the volant animal exclusion treatment compared to the control (P<0.05), but there was no difference between the climbing animal exclusion treatment and the control (P>0.05). The proportions of fruits removed at night was significantly higher than during day time (P<0.001, Fig. 4.1) in the control and in the climbing animal exclusion treatment, but not in the volant animal exclusion treatment (P>0.05). Proportions of fruits removed during day time were 25% in the control, 19% in the climbing animal exclusion and zero in the volant animal exclusion treatments, and without significant differences between each other (t test, P>0.05). The frugivorous bird species take away almost 19% mature fruits during day time. Bats at night play a more important role than birds at day on the primary seed dispersal of wild banana (t test, P<0.05). Removed fruits (%)
100
80
Fruits removed at night
Fruits removed in daytime
a
a
60
40
b
20
0
Open control
b
CA excluded
Treatment
b
b
VA excluded
Fig. 4.1 Proportion of fruits removed from wild banana plants in different treatments at night and during the day. Data are means ± SE, n=8 plants per treatment. Bars with different letters are significantly different at P<0.001. This experiment was conducted only at Site 1 (XTBG, a disturbed fragment of a tropical seasonal lowland rain forest). CA and VA means climbing animals and volant animals, respectively. 72 Chapter 4 Seed dispersal of Musa acuminata 4.3.2. Secondary seed dispersal and predation In total, the proportions of seeds that remaining in the seed removal experiments of all sites and habitats were 14% in the control, 31% in the ant exclusion and 44% in the rodent exclusion treatments. The ant exclusion and the rodent exclusion treatments represented a significant difference between each other (P<0.05, Fig. 4.2). 50
Remaining seeds (%)
40
aa
Control
Ants excluded
Rodents excluded
bb
30
20
c
10
0
Treatment
Fig. 4.2 Proportions of experimentally exposed wild banana seeds that remaining in different treatments (control, exclusion of ant and exclusion of rodent). Data are means ± SE, 630 seeds per treatment. Bars with different letters are significantly different at P<0.05. In the control treatment, all combined influences of season, site and habitat significantly affected the secondary seed removal of wild banana, with the exception of interactions of season/site and season/habitat (Table 4.1). The proportion of remaining seeds was 27% in the rainy season and 13.8% in the dry season. The differences were significant at P<0.05 (Fig. 4.3a). The proportions of remaining seeds per site were 17.5% at Site 1, 19.5% at Site 2 and 26.3% Site 3 (Fig. 4.3b) and were significantly higher at Site 3 than at the other sites (P<0.05). Proportions of remaining seeds were significantly different between all the habitat types with the highest proportion in forest without wild banana plants (29%) and lowest in forest with wild banana stands (14%; Fig. 4.3c) (P<0.05). In the habitats, the highest proportion of seed remaining occurred in the forest without bananas at Site 3 (37%), followed by forest without bananas at Site 2 (32%) and the open land at Site 3 (30%). The proportion of remaining seeds in these three habitats was significantly higher compared to all other habitats (P<0.05; Fig. 4.3d). In the ant exclusion treatments, the only factor which was not significant was the 73
Chapter 4 Seed dispersal of Musa acuminata a
35
Remaining seeds (%)
a
25
20
b
15
10
5
0
Rainy
35
Remaining seeds (%)
30
25
20
15
c
Season
b
c
5
M1
M2
Habitat
b
a
25
M3
b
b
20
15
10
5
50
a
10
0
30
0
Dry
Remaining seeds (%)
Remaining seeds (%)
35
30
S1
S2
S3
Site
d
a
40
a
a
30
20
b
b
b
b
b
b
10
0
1 2 3
1
3
1 2 3
2
M M M M M M M M M
S1 S1 S1 S2 S2 S2 S3 S3 S3
Site and habitat
Fig. 4.3 Proportions of experimentally exposed wild banana seeds that remaining in the control treatments: (a) in the dry and in the rainy season (ndry season=360 seeds totally exposed, nrainy season=270 seeds totally exposed), (b) at different sites (n=210 seeds per site), (c) in different habitat types (n=210 seeds per habitat type) and (d) in all habitats (n=70 seeds per habitat). Data are means ± SE, bars with different letters are significantly different at P<0.05. S represents Sites (1, 2 and 3 including all habitats), M represents habitat type (1 = forest with wild banana stands, 2 = forest without wild banana, 3 = open land without wild banana), it’s same with Fig. 4.4 and Fig. 4.5. effect of season (Table 4.1). The proportion of seeds that remaining at Site 1, Site 2 and Site 3 was 38%, 32.5% and 50% respectively and was significantly higher at Site 3 compared to the two other sites (P<0.05; Fig. 4.4a). In forest with wild banana stands, 30% of seeds remaining and which was a significantly lower proportion than in forest without bananas (47.5%) and open land (41.75%, P<0.05; Fig. 4.4b). All factors had significant effects in the rodent exclusion treatments (Table 4.1). The proportion of seeds that remaining at Site 3 was 61% and significantly higher than at Site 1 (47%) and Site 2 (45.2%, P<0.05; Fig. 4.5a). The latter two sites showed no significant differences between each other. The rate of seeds that remaining in forest with wild banana stands was 61% and significantly higher than in forest without bananas (47%) and in open land (43%) ( P<0.05; Fig. 4.5b). 74 Chapter 4 Seed dispersal of Musa acuminata 70
60
70
a
a
50
b
40
b
30
20
10
0
S1
S2
Site
S3
Remaining seeds (%)
Remaining seeds (%)
60
b
a
50
40
a
b
30
20
10
0
M1
M2
Habitat
M3
Fig. 4.4 Proportions of remaining wild banana seeds in the ant exclusion treatment in (a) different sites and (b) different habitat types (n= 210 seeds in each category of a and b) Data are means ± SE, bars with different letters are significantly different at P<0.05. 60
50
a
a
b
b
40
30
20
10
0
S1
S2
Site
S3
70
Remaining seeds (%)
70
Remaining seeds (%)
60
b
a
b
50
b
40
30
20
10
0
M1
M2
Habitat
M3
Fig. 4.5 Proportion of remaining wild banana seeds in the rodent exclusion treatment in (a) different sites and (b) different habitat types (n= 210 seeds in each category of a and b). Data are means ± SE, bars with different letters are significantly different at P<0.05. 4.3.3. Wild banana seed dispersers and predators Observation during daytime showed that the black‐headed bulbul Pycnonotus atriceps (Pycnonotidae) is the main species which take away almost 19% mature fruits during day time. Additionally, the white eye Zosterops palpebrosus (Zosteropidae) and the little spider hunter Arachnothera longirostra (Nectariniidae) are the main birds species visiting inflorescences of wild banana and peck some pulp. The short‐nosed fruit bat (Cynopterus sphinx) (Pteropodidae) was observed to remove fruits of wild banana at night. The species begins to visit fruit‐bearing plants about 30 min after sunset. Nine species of rodents and totally 97 individuals were captured in all habitats and sites (Table 4.2). The four dominant species were Rattus tanezumi, Apodemus 75
Chapter 4 Seed dispersal of Musa acuminata confucianus, Rattus fulvescens and Maxomys surifer belonging to family of Muridae and which accounted for 36%, 24%, 18% and 11% respectively of the total number of individuals. Site 3 had the highest number of rodent species and individuals (7 species and 35 individuals). Wild banana stands in forest had the highest number of rodent species and individuals (8 species and 48 individuals) (Table 4.2). Table 4.1 Multivariate analysis of the effects of season, site and habitat on the secondary removal of wild banana seeds in the control, ant exclusion and rodent exclusion treatments Control Season Site Habitat Season*site Season*habitat Site*habitat Season*site*habitat Error Total Ant exclusion Season Site Habitat Season*site Season*habitat Site*habitat Season*site*habitat Error Total Rodent exclusion Season Site Habitat Season*site Season*habitat Site*habitat Season*site*habitat Error Total ***: P<0.001; **: P<0.01; ns: P>0.05 df MS F P ‐value 1 2 2 2 2 4 4 612 630 1 2 2 2 2 4 4 612 630 1 2 2 2 2 4 4 612 630 274.667 47.114 111.905 0.879 11.804 43.776 36.703 8.779 35.493 180.322 188.358 170.986 47.431 196.191 115.888 11.407 29.484 115.053 167.416 113.519 43.911 29.142 155.547 11.910 31.286 5.366 12.746 0.100 1.344 4.986 4.181 3.112 15.809 16.513 14.990 4.158 17.200 10.160 2.506 9.660 14.057 9.531 3.687 2.447 13.060 0.000*** 0.005** 0.000*** 0.905ns 0.261ns 0.001*** 0.002** 0.078ns 0.000*** 0.000*** 0.000*** 0.016* 0.000*** 0.000*** 0.039* 0.000*** 0.000*** 0.000*** 0.026* 0.045* 0.000*** Totally 104 species and 2,600 individuals of ants were captured in all habitats and sites (Table 4.2). The three dominant species were Pheidole yeensis, Pheidole rhombinoda and Odontoponera transversa belonging to family of Formicidae and which accounted for 29%, 11% and 10% respectively of the total number of individuals. Pheidole sinensis and Pachycondyla luteipes were the subdominant 76 Chapter 4 Seed dispersal of Musa acuminata species which accounted for 7% and 4% respectively. Site 2 had the highest number of ant species and individuals (64 species and 1,310 individuals). Open land habitats had the highest number of ant species and individuals (61 species and 1,698 individuals) (Table 4.2). Table 4.2 The number of species and individuals of rodents and ants in different season, site and habitat. Group
Numbers
Dry
Rainy S1*
S2
S3 M1*
M2
M3
Total
Species
8
5
6
5
7
8
4
4
9
Individuals
47
50
32
30
35
48
42
7
97
Species
72
53
46
64
49
41
49
61
104
Individuals
1778
822
469 1310 821 491
411
Rats
Ants
1698 2600
*S represents Sites (1, 2 and 3 including all habitats), *M represents habitat type (1 = forest with wild banana stands, 2 = forest without wild banana, 3 = open land without wild banana). 4.4 Discussion From the high fruit removal rate at night‐time of about 80%, it can be concluded that most of the fruits were removed by bats. Previous studies (Tang et al. 2007) confirmed that Cynopterus sphinx is the most important volant mammal consuming wild banana fruits in the study region. The bats seldom remain on the plants to feed but instead carry fruits to feeding roosts, repeating this behavior several times throughout the night (Elangovan et al. 1999; Tang et al. 2007). During the flights from the foraging plants to the roosts, some fruits usually get lost. Because the transportation distance of fruits by bats reaches 100‐1,000 m (Corlett 2009) banana seed may be dispersed to various other habitats besides the refuse piles near the roost. The loss of bat species will take negative effects on the expansion of wild banana population to uncolonized space. Rodents accounted for a significantly higher total seed removal rate than ants, and therefore destroyed most of the dispersed seeds. The proportion of seeds removed by rodents was significantly higher in the forest with wild banana stands compared to the latter two habitats, indicating that the highest rate of seed predation occurred in 77
Chapter 4 Seed dispersal of Musa acuminata the vicinity of the parent plants. Although the use of the seed removal rate may not be a reliable proxy for seed predation (Vander Wall et al. 2005a; Cole 2009), amounts of seed debris close to the Petri dishes indicate that large proportions of seeds were consumed by rodents directly after detection. Accordingly, other studies also showed that seed predation by small rodents was the major limitation for seedling establishment during the process of seed dispersal (Bricker et al. 2010), and seed predation was also found to explain the landscape‐level abundance in another early‐successional plant (Phytolacca americana; Orrock et al. 2006). However, it was also found that small rodents can contribute to seed dispersal by their scatter‐hoarding behavior (Brewer & Rejmánek 1999). The high rodent density and seed predation rates at forest sites with wild banana can be explained by a dense vegetation structure, providing shelter and protection from predators, in addition to the provision of a rich food source, corresponding with the conclusions drawn by Hulme and Kollmann (2005). The numbers of rodent species and individuals recorded from these three sites XTBG, Xinshan and Bubeng were nearly identical but the lowest proportion of seeds was removed from Bubeng which is well protected and relatively undisturbed so the structure is not as dense and there is little shelter for the rodents. Fedriani and Manzaneda (2005) also found that seeds located in sheltered microsites experienced higher removal by seed predators compared with seeds in open microsites. Therefore, spatial patterns of seed predation are not only affected by the rodent population densities, but also by other factors such as habitat type (Calvĩno‐Cancela & Martín‐Herrero 2009), the degree of disturbance or habitat degradation (Cole 2009), seed availability and habitat physiognomy (Díaz 1992), predator’s preferences (Vanhoenacker et al. 2009), habitat preferences of predators (Manson & Stiles 1998) or edge effects (Notman et al. 1996). Furthermore, the high seed predation rate in the wild banana stands may reduce competition between seedlings and the parent plants, according to the escape hypothesis of Janzen‐Connell (Janzen 1970; Connell 1971). Ants in this study also played role on the secondary seed dispersal for wild banana to some extent. Studies on the rain forest herbs of the genus Globba (Zingiberaceae) showed maximum dispersal distances by ants of 8 m in Borneo (Pfeiffer et al. 2004) and 3.3 m in China (Zhou et al. 2007). Another effect of ants on seeds was found in Ficus benjamina, showing that the seeds treated by ants gained a significantly higher germination rate, although the dispersal distance by ants was only 1.8 m (Zhang & Chen 2008). A general effect of seed dispersal by ants is the scattering of aggregated 78 Chapter 4 Seed dispersal of Musa acuminata seeds, as shown by Chen et al. (2004) from the seed dispersal of Globba lancangensis. We observed that ants in this study can move the seeds 1‐2m away from the Petri dishes. Volant animals and ants provide complementary seed dispersal at different spatial scales. By acting as secondary dispersers, ants may also provide a fine‐tuned dispersal following long‐distance dispersal by volant animals (Christianini & Oliveira 2010). Ness and Morin (2008) point out that ant‐dispersed plants are often conspicuously rare near forest edges relative to forest interiors. In the present study, the open land adjacent to forest had the highest species richness and abundance of ants and the highest seed removal rate by ants. However, considering the different sites, there was no close correlation between species diversity and seed removal rate. This may be explained by different preferences of the various ant species for wild banana seeds. The seed removal rate by ants in the dry season was significantly higher than in the rainy season, which is in accordance with the observation of Basu (1997) who found that ant response to humidity and that most ant species are more active in the dry season. This seasonal pattern provides an advantage for the wild banana population because the seeds are distributed in the dry season and increases the chance of successful germination and seedling establishment in the rainy season. In conclusion, the results of this study showed strong relationships between wild banana seeds and different animals, acting as seed dispersers or seed predators. Primary seed dispersal showed a high dependency from bats and can be considered as the most important primary seed disperser. Wild banana stands has the most abundant rodents and the highest seed predation and the lowest proportion of secondary seed dispersal by ants. In contrast, seed predation by rodents in forest and open habitat was relatively weak and a large proportion of post dispersal seeds were removed by ants. Therefore, primary seed dispersal mostly by bats at night time is essential for wild banana seeds to escape seed predation. 4.5 References Baker HG (1974). The evolution of weeds. Annual Review of Ecology and Systematics 5, 1‐24. Basu P (1997). Competition hierarchy in the ground foraging ant community in a wet evergreen forest (Western Ghats, India): Role of interference behaviour. Current 79
Chapter 4 Seed dispersal of Musa acuminata Science 73, 173‐179. Brewer SW (2001). Predation and dispersal of large and small seeds of a tropical palm. Oikos 92, 245‐255. Brewer SW, Rejmánek M (1999). Small rodents as significant dispersers of tree seeds in a Neotropical forest. Journal of Vegetation Science 10, 165‐174. Bricker M, Pearson D, Maron J (2010). Small‐mammal seed predation limits the recruitment and abundance of two perennial grassland forbs. Ecology 91, 85‐92. Calvĩno‐Cancela M, Martín‐Herrero J (2009). Effectiveness of a varied assemblage of seed dispersers of a fleshy‐fruited plant. Ecology 90, 3503‐3515. Cao M, Zou X, Warren M, Zhu H (2006). Tropical forests of Xishuangbanna, China. Biotropica 38, 306‐309. Chen F, Chen J, Liu ZQ, Zhang L, Liu Y, Bai ZL (2004). The role of ants in seed dispersal of Globba lancangensis and the spatial distribution of its seedlings. Acta Phytoecological Sinica 28, 210‐217. (in Chinese with English abstract) Christianini AV, Oliveira PS (2010). Birds and ants provide complementary seed dispersal in a neotropical savanna. Journal of Ecology 98, 573‐582. Connell JH (1971). On the role of natural enemies in preventing competitive exclusion in some marine animals and in rain forest trees. In: den Boer BJ and Gradwell GR, ed. Dynamics of population. Centre for Agricultural Publishing and Documentation, Wageningen, pp. 298‐310. Corlett RT (2009). Seed Dispersal Distances and Plant Migration Potential in Tropical East Asia. Biotropica 41, 592‐598. Cole RJ (2009). Postdispersal Seed Fate of Tropical Montane Trees in an Agricultural Landscape, Southern Costa Rica. Biotropica 41, 319‐327. Davidson DW, Morton SR (1981). Myrmecochory in some plants (F. Chenopodiaceae) of the Australian arid zone. Oecologia 50, 357‐366. Díaz M (1992). Spatial and temporal patterns of granivorous ant seed predation in pathy cereal crop areas of central Spain. Oecologia 91, 561‐568. Elangovan V, Marimuthu G, Kunz TH (1999). Temporal patterns of individual and group foraging behaviour in the short‐nosed fruit bat, Cynopterus sphinx, in south India. Journal of Tropical Ecology 15, 681‐687. Fedriani JM, Manzaneda AJ (2005). Pre‐ and postdispersal seed predation by rodents: balance of food and safety. Behavioral Ecology 16, 1018‐1024. Forget PM, Milleron T (1991). Evidence for secondary seed dispersal by rodents in panama. Oecologia 87, 596‐599. 80 Chapter 4 Seed dispersal of Musa acuminata García‐Castaño JL, Kollmann J, Jordano P (2006). Spatial variation of post‐dispersal seed removal by rodents in highland microhabitats of Spain and Switzerland. Seed Science Research 16, 213‐222. Herrera CM (2002). Seed dispersal by vertebrates. In: Herrera CM and Pellmyr O, ed. Plant–animal Interactions: An Evolutionary Approach. Blackwell Science, Oxford, pp. 185–210. Hulme PE (1994). Postdispersal seed predation in grassland ‐ its magnitude and sources of variation. Journal of Ecology 82, 645‐652. Hulme PE, Kollmann J (2005). Seed predator guilds, spatial variation in post‐dispersal seed predation and potential effects on plant demography‐ a temperate perspective. In: Forget PM, Lambert JE, Hulme PE, Vander Wall SB, ed. Seed fate: predation, dispersal and seedling establishment. CABI Publishing, Wallingford, pp. 9‐30. Janzen DH (1970). Herbivores and the number of tree species in tropical forests. The American Naturalist 104, 501‐528. Jordano P (1992). Fruits and frugivory. In: Fenner M, ed. Seeds: The Ecology of Regeneration in Plant Communities. CAB International, Wallingford, UK, pp. 105‐156. Kolb A, Ehrlén J, Eriksson O (2007). Ecological and evolutionary consequences of spatial and temporal variation in pre‐dispersal seed predation. Perspectives in Plant Ecology Evolution and Systematics 9, 79‐100. Levey DJ Byrne MM (1993). Complex ant plant interactions ‐ rain‐forest ants as secondary dispersers and postdispersal seed predators. Ecology 74, 1802‐1812. Liu AZ (2001). Phylogeny and Biogeography of Musaceae, Ph.D. Dissertation. Kunming Institute of Botany, The Chinese Academy of Sciences, Kunming, Yunnan, P.R. China. (in Chinese with English abstract) Lü XT, Yin JX, Tang JW (2010). Structure, tree species diversity and composition of tropical seasonal rainforests in Xishuangbanna, south‐west China. Journal of Tropical Forest Science 22, 260‐270. Manson RH, Stiles EW (1998). Links between microhabitat preferences and seed predation by small mammals in old fields. Oikos 82, 37‐50. Medellín RA, Gaona O (1999). Seed dispersal by bats and birds in forest and disturbed habitats of Chiapas, Mexico. Biotropica 31, 478‐485. Myster RW (2003). Effects of species, density, patch‐type, and season on post‐dispersal seed predation in a Puerto Rican pasture. Biotropica 35, 542‐546. Notman E, Gorchov DL, Cornejo F (1996). Effect of distance, aggregation, and habitat on levels of seed predation for two mammal‐dispersed neotropical rain forest tree 81
Chapter 4 Seed dispersal of Musa acuminata species. Oecologia 106, 221‐227. Ness JH, Morin DF (2008). Forest edges and landscape history shape interactions between plants, seed‐dispersing ants and seed predators. Biological Conservation 141, 838‐847. Orrock JL, Levey DJ, Danielson BJ, Damschen EI (2006). Seed predation, not seed dispersal, explains the landscape‐level abundance of an early‐successional plant. Journal of Ecology 94, 838‐845. Pfeiffer M, Nais J, Linsenmair KE (2004). Myrmecochory in the Zingiberaceae: seed removal of Globba franciscii and G. propinpua by ants (Hymenoptera‐Formicidae) in rain forests on Borneo. Journal of Tropical Ecology 20, 705‐708. Rodríguez‐Pérez J, Traveset A (2010). Seed dispersal effectiveness in a plant‐lizard interaction and its consequences for plant regeneration after disperser loss. Plant Ecology 207, 269‐280. Ruiz J, Boucher DH, Chaves LF et al. (2010). Ecological consequences of primary and secondary seed dispersal on seed and seedling fate of Diptetyx oleilera (Fabaceae). Revista De Biologia Tropical 58, 991‐1007. Schupp EW, Jordano P, Gómez JM (2010). Seed dispersal effectiveness revisited: a conceptual review. New Phytologist 188, 333‐353. Shepherd VE, Chapman CA (1998). Dung beetles as secondary seed dispersers: impact on seed predation and germination. Journal of Tropical Ecology 14, 199‐215. Shugart HH (1998). Terrestrial Ecosystems in Changing Environments. Cambridge University Press, Cambridge. pp.103‐143. Sivy KJ, Ostoja SM, Schupp EW, Durham S (2011). Effects of rodent species, seed species, and predator cues on seed fate. Acta Oecologica 37, 321‐328. Tang, ZH, Sheng LX, Ma XF et al. (2007). Temporal and spatial patterns of seed dispersal of Musa acuminata by Cynopterus sphinx. Acta Chiropterologica 9, 229‐235. Thomas DW, Cloutier D, Provencher M, Houle C (1988). The shape of bird‐generated and bat‐generated seed shadows around a tropical fruiting tree. Biotropica 20, 347‐348. Vander Wall SB, Kuhn KM, Beck MJ (2005b). Seed removal, seed predation, and secondary dispersal. Ecology 86, 801‐806. Vander Wall SB, Kuhn, KM, Gworek JR (2005a). Two‐phase seed dispersal: linking the effects of frugivorous birds and seed‐caching rodents. Oecologia 145, 282‐287. Vander Wall SB, Longland WS (2004). Diplochory: are two seed dispersers better than one? Trends in Ecology & Evolution 19, 155‐161. Vanhoenacker D, Ågren J, Ehrlén J (2009). Spatial variability in seed predation in Primula 82 Chapter 4 Seed dispersal of Musa acuminata farinosa: local population legacy versus patch selection. Oecologia 160, 77‐86. von Allmen C, Morellato LPC, Pizo MA (2004). Seed predation under high seed density condition: the palm Euterpe edulis in the Brazilian Atlantic Forest. Journal of Tropical Ecology 20, 471‐474. Wang BC, Smith TB (2002). Closing the seed dispersal loop. Trends in Ecology & Evolution 17, 379‐385. Zhang S, Chen J (2008). Secondary seed dispersal of Ficus benjamina, New evidence for ant non‐myrmecochorous mutualism. Chinese Journal of Ecology 27, 1913‐1919. (in Chinese with English abstract). Zhou HP, Chen J, Chen F (2007). Ant‐mediated seed dispersal contributes to the local spatial pattern and genetic structure of Globba lanciangensis (Zingiberaceae). Journal of Heredity 98, 317‐324. 83
Chapter 4 Seed dispersal of Musa acuminata 84 Chapter 5 The role of EFNs & ants on Leea Chapter 5 Young leaf protection in the shrub Leea glabra in south–west China: the role of extrafloral nectaries and ants Meng, L.‐Z., Martin K., Liu, J.‐X., Chen, J. (2011). Young leaf protection in the shrub Leea glabra in south‐west China: the role of extrafloral nectaries and ants. Arthropod‐Plant Interactions, 1‐7. DOI:10.1007/s11829‐011‐9151‐6. 85
Chapter 5 The role of EFNs & ants on Leea Abstract Field experiments on Leea glabra in its natural forest habitat of southern Yunnan, China were conducted to study the effects of artificial damage of young and old leaves on extrafloral nectaries (EFNs) secretion quantity and sugar concentration, as well as the effects on ant abundance on the plants following the damage treatments. We found there were no rapid changes in extrafloral nectar volume or nectar sugar concentration which would indicate an induced reaction following artificial damage. However, both cutting and punching of young leaves resulted in a significant increase (2–4‐fold) of ants within 6 h after damage compared to undamaged controls. In another experiment, disks of fresh young L. glabra leaves that were pinned on young leaves of another L.glabra plant also resulted in a significant increase in the number of ants compared to treatment with paper disks, indicating that ants were most probably attracted by volatile organic compounds (VOCs) released from damaged young leaves. Furthermore, we found that portion of damaged leaf area of young leaves was significantly lower than that of old leaves and the concentration of tannins was significantly higher in young than in medium and old leaves. In conclusion, our results show that young leaves of L. glabra are protected against attacks by herbivores by multiple mechanisms, which include: (1) the activity of EFNs, which attract different ant species from the surrounding ground; (2) a mechanism induced by the damage of young leaves, which leads to rapidly increased ant recruitment and is most probably caused by the release of volatiles from damaged leaf and (3) a higher allocation of tannins in young than in older leaves. Keywords: Ant–plant interactions; Leea glabra; Plant cues; Simulated herbivory; Tannin; VOCs 86 Chapter 5 The role of EFNs & ants on Leea 5.1 Introduction Extrafloral nectaries (EFNs) are found in more than 90 plant families and about 300 genera. They have a high diversity of shapes, patterns of distribution on plants and secretion products (Koptur 1992). EFNs are generally considered as structures established to attract defending arthropods by providing nectar as a food resource, especially for ants (Heil 2008). Based on a meta‐analysis of ant–plant protection mutualisms, Chamberlain and Holland (2009) concluded that ant effects on plants are routinely positive for plants, and only occasionally neutral. The most pronounced effects of ants as biotic defenses are from tropical systems and for true myrmecophytic plants, even though ants consume extrafloral nectar when no herbivores are present, decreasing plant fitness (Rosumek et al. 2009). An interesting question is whether the quantity or quality of extrafloral nectar production can be increased or improved following damage in order to increase patrolling by defending ants. This issue was reviewed by Agrawal and Rutter (1998) who showed an increase in extrafloral nectar volume and the number of EFNs in different EFN‐plant species after artificial plant damage, indicating a mechanism of induced defense. However, there are also studies which showed no significant effects of damage on nectar production or quality (Koptur 1989; Smith et al. 1990). Other studies demonstrated that ants recruit specifically to damaged leaves and respond rapidly to herbivores in obligate ant–plant systems (Agrawal 1998a; Agrawal & Dubin‐Thaler 1999; Grangier et al. 2008; Romero & Izzo 2004), but the potential cues for the rapid recruitment of ants were not always clear due to the few studies that measured both ant activity in response to damage and nectar flow in the field. Furthermore, a question remains as to whether constitutive or induced defense mechanisms differ between plant tissues in relation to the value of the tissue for the plant, i.e., whether plants invest more defenses in higher value parts (e.g., young leaves) than in those of lower value (e.g., old leaves). Such conditions are discussed in the optimal defense hypothesis, with the basic assumption that defense is costly for plants and therefore deployment among tissues is directly related to their value and likelihood of herbivore attack (McKey 1979; Rhoades 1979). In addition to the attraction of defending arthropods to deter herbivory, many plant species with EFNs employ a diverse array of other chemical and physical defenses (Agrawal & Rutter 1998), especially the production of chemical compounds such as tannins and phenolics (Coley 1986). Young leaves of various plant species are often 87
Chapter 5 The role of EFNs & ants on Leea endowed with higher concentrations of secondary metabolites which form a combined defensive strategy for the adaptation of selective pressure from different herbivore communities (Coley et al. 2005; Raupp & Denno 1983). The present study deals with the shrub Leea glabra and the defensive function of the EFNs found at leaf petioles in relation to their interactions with ants under different conditions. Field experiments with naturally grown L. glabra plants were conducted in a tropical forest plot of southern Yunnan (south–west China) to answer the following major questions: (1) is the quantity of EFN secretions and the sugar concentration of the nectar in artificially damaged young leaves higher than in old leaves? (2) Are ant abundances higher after damage on young leaves than damage on old leaves? If so, is ant recruitment related to changes in EFN production or to the release of VOCs after artificial damage? (3) Are there differences in tannin concentrations in leaves of different ages, indicating tannins are an additional mechanism of herbivore defense? 5.2 Methods 5.2.1 Study site The study area is located in the Naban River Watershed National Nature Reserve (NRWNNR) within the Dai autonomous prefecture of Xishuangbanna, southern Yunnan Province, south–west China (22°10′ N, 100°38′ E). The climate is humid northern marginal tropical monsoonal with three distinct seasons: cool‐dry (October–January), hot‐dry (February–May) and a rainy season (June–September) when most of the mean annual precipitation of almost 1600 mm occurs. The specific study site was a middle‐slope forest area located at about 900 m asl. It is a disturbed fragment of a lower hill tropical seasonal rainforest with a maximum tree height of about 30 m. The forest structure is characterized by two to three strata, and Pometia tomentosa, Bauhinia variegata and Kydia calycina are the dominant tree species. The understorey is characterized by Litsea monopetala and the herb Curculigo capitulata together with the seedlings of canopy tree species. All plant samples observed in this study were growing under roughly uniform soil and light conditions. 88 Chapter 5 The role of EFNs & ants on Leea 5.2.2 Study species Leea glabra C. L. Li (Leeaceae: Li 1996) is an erect shrub species with a height of 1.5–3.5 m and is one of about 70 Leea species. It is native to the southern parts of Yunnan and Guangxi Provinces of south–west China and is mainly distributed in the lower layer of forests up to 1200 m asl. (Flora of China, eFloras.org). In our study site, about 500 L. glabra plants were distributed over an area of 5 ha. According to our field observations on L. glabra in the study site over a period of one year (May 2008–April 2009), the plants produce flowers only once per year between April and May before the rainy season. Fruits are produced in August at the end of the rainy season. Production of new leaves was observed throughout the year on branches already bearing old leaves. Single units of extrafloral nectaries (EFNs) of four to six parallel strips not exceeding 1 cm in length are developed on the stem on the opposite side of the buds of new leaves or flowers. When the young leaves are mature and change their color from red to green, the EFNs become inactive and stop nectar secretion. It generally takes young leaves almost four weeks to mature. A lepidopteran larval is the major herbivore of mature leaves on which it produces many clear holes in parallel strips. We observed about 10 species of ants assembling at the EFNs with Camponotus singularis, Oecophylla smaragdina and Crematogaster rothneyi being the most common, but usually there was only one species of ant present at a time on one plant. The ants do no nest on the plant, but visit the plant from the surrounding ground, patrolling the leaves and consuming nectar from the EFNs. Ants were never found on inflorescences during the flowering period of about five days. Experiment 1: EFN measurements To test the effects of simulated herbivore damage on EFN production and sugar content, field experiments were conducted in April 2009 before the flowering period. A total of 20 plants that were almost three years old and approximately 2.5 m tall were selected randomly for experiments. All plants were individually enclosed in white nylon nets (mesh size 0.5 mm), supported by woody sticks to avoid contact with the plant: this was done to protect plants from herbivore attacks. Each selected experimental plant stood at least 10 m apart from the next. Because all the experimental plants were covered from the top down but the nylon net did not touch the ground, it should not have influenced ant activity. 89
Chapter 5 The role of EFNs & ants on Leea To study the effects of different types of herbivore damage on extrafloral nectar quantity and sugar concentration, we surrounded the active EFNs with grease on one young or one old leaf with no signs of previous herbivore damage and artificially damaged the leaves by either cutting or punching. Cutting was conducted by removing the distal half (50%) of each leaflet using scissors. Punching was conducted by removing 50% of the each leaflet area over the whole leaf surface using a 3‐mm‐diameter hole‐punch. These methods of artificial leaf damage were adapted from Pulice and Packer (2008). The two damage methods mimic the most common natural damage form found on the plants. In total, the following treatments to net‐covered plants were conducted: (1) cutting young leaves; (2) cutting old leaves; (3) punching young leaves and (4) punching old leaves. Each treatment was applied to four plants, and four undamaged net‐covered plants served as controls. The young leaves used for damage experiments were two weeks past leaf bud development, and old leaves were six weeks past the leaf color change from red to dark green. The old leaves selected for treatments were adjacent to a young leaf, sharing the same branch with the EFNs selected for nectar measurements. Quantitative nectar measurements after damage by cutting or punching were conducted in the following way: the total volume of nectar produced by the EFN units of the branch was collected using micro‐capillaries with a minimum scale of 0.5 μL (Abel Industries Inc., USA). Nectar was drawn into the micro‐capillary tubes by means of capillary action. In addition to nectar quantity, the sugar concentration in Brix units was measured by a refractometer (B+S Instruments, UK). The method used to measure the sugar concentration within nectar was adapted from Heil et al. (2000). The grease did not interfere with the nectar measurements. Measurements were conducted prior to damage, then 10 min after damage, and followed by 10 further records at time intervals of 6 h for a total of 12 records within a period of almost 60 h. The same measurements were also made on the four control plants on the same days. Experiment 2: Ant behavior Effects of simulated herbivore damage on ant recruitment to young L. glabra leaves were analyzed in June 2010 by counting numbers of ants on artificially damaged and undamaged young and old leaves of L. glabra. The following treatments were conducted on selected plants with both young and old leaves present and without signs of recent herbivore damage: young leaves of 20 plants were damaged by 90 Chapter 5 The role of EFNs & ants on Leea removing the distal half of a leaf by cutting, and 20 plants were damaged by removing half of the leaf area by hole punching, as described for the previous experiment. The same two types of damage were also applied to old leaves using 20 separate plants each. Each selected plant stood at least 10 m apart from the next. A total of 20 undamaged plants located 50 m from the plot of damaged plants were used as controls for the respective treatments. All plants observed were of a similar age, almost three years old. Numbers of ants on the leaves of the experimental and the control plants were counted prior to treatment, and 10 min, 6 and 24 h after treatment. All treatment and control plants were monitored during the same period, so there would not be temporal differences in ant activity. To look for an indication that ants might be attracted by VOCs released by wounded plants (Agrawal & Dubin‐Thaler 1999), a further experiment was conducted: six disks of freshly cut young L. glabra leaves of 6 cm2 each were pinned randomly on leaflets of one young undamaged compound leaf of another L. glabra plant. Leaf disks were from another three‐year‐old plant. These treatments were conducted on 20 plants, and another 20 control plants were located in another plot which stood 50 m from the plot of treatment plants. The controls were treated with paper disks applied in the same way. Experiment 3. Tannin and leaf damage measurements To test for a biochemical difference between leaves of different development stages, leaf samples of three age categories were collected and analyzed for tannin concentration: (1) young leaves, (2 week old); (2) medium‐aged leaves (4 week old); and (3) old leaves (8 week old). The three categories of leaves were collected from the same plant at same time from a total of 12 plants randomly selected. Damage to leaves at different development stages was measured using graph paper in the field to calculate the damaged leaf area ratio of leaves at different ages. All leave samples used were fully expanded at the time of measurement. Total tannin concentrations were then determined colorimetrically by spectrophotometric techniques in the laboratory. The Folin–Ciocalteau (Laborclin) method was used to determine the tannin content of the leaf samples. First, calibration curves were made with different concentrations of the pyrogallol standard. Then, a water coloration reagent (Folin– Ciocalteau) and sodium carbonate solution were proportionally added to a prepared extract solution of crushed leaf powder in ethanol. Finally, the blended solution was shaken and its optical density measured at 760 nm after coloration for 30 min. 91
Chapter 5 The role of EFNs & ants on Leea 5.2.3 Statistical analysis All data were analyzed using SPSS statistical software 13.0 version (SPSS Inc. 2004). To ensure that there were no differences among the treatment groups of plants prior to damage, initial plant height and number of leaf twig layers of plants used for the EFN production and sugar concentration measurements were compared with one‐way ANOVA. Repeated measures ANOVA procedures were used to compare EFN production and sugar concentration between different treatments and controls over time. In these procedures, time was considered as the factor of repetition, and treatments and plants were considered fixed and random effects, respectively. Then post hoc tests were used for the paired comparison between different treatments on young and old leaf with a Bonferroni correction. Differences among the number of ants after treatments were analyzed by repeated measures ANOVA followed by Dunnett’s contrast with a Bonferroni correction also. To account for violations of the sphericity assumption of variances, the Greenhouse–Geisser (G–G) and Huynh–Feldt (H–F) correction was applied to the degrees of freedom based on the ε value which is more or less than 0.75 produced by Mauchly’s test before (Zar 1996). Tannin concentrations and the damaged leaf area ratio of leaves at different ages were compared using MANOVA with leaf age as the independent variable and the plant was included as a random factor. 5.3 Results EFNs and sugar Nectar production did not differ significantly between the four different treatments and controls within 60 h after simulated damage (Fig. 5.1a; repeated measures ANOVA: F4, 15=0.971, P = 0.452). No significant effects on EFN production over time were found in comparison to control plants after the G‐G correction (Fig. 5.1a; repeated measures ANOVA: F3.13, 46.90=2.73, P = 0.052). Plants differed within the young leaf cutting treatment (Fig. 5.1a; P < 0.01). However, this among plant variation was not observed within the three other types of treatment (Fig. 5.1a; P > 0.05). There was no significant effect of leaf age on nectar production for either damage treatment (paired comparison with a Bonferroni correction; Fig. 5.1a; P > 0.05). Sugar concentrations did not differed significantly between the four treatments and controls within 60 h after simulated damage (Fig. 5.1b; repeated measures ANOVA: F4, 15=1.26, 92 P = 0.273). No significant effects on sugar concentration over time were Chapter 5 The role of EFNs & ants on Leea found in comparison to control plants after the G‐G correction (Fig. 5.1b; repeated measures ANOVA: F4.26, 63.86=1.05, P = 0.187). Nectar quality also did not differ between the young and old leaves following either damage treatment (Fig. 5.1b; P > 0.05, with Bonferroni correction). 12
30
a (n=4)
CY
CO
PY
PO
CK
10
EFN production (uL)
8
6
4
2
Sugar concentration percentage(%)
bb (n = 4)
25
20
15
10
5
0
0
0
10min
6h
12h
18h
24h
30h
36h
42h
48h
54h
60h
Time after treatment
0
10min
6h
12h
18h
24h
30h
36h
42h
48h
54h
60h
Time after treatment
Fig. 5.1 Mean values of EFN production (a) and sugar concentration percentage (b) within nectar after damage treatment on young and old L. glabra leaves by cutting and punching at different times compared to undamaged control leaves. t0 corresponds to the values immediately prior to damage (CY, CO, PY, PO and CK represent four treatments of cutting young leaves, cutting old leaves, punching young leaves, punching old leaves and control respectively, and it is same with Fig. 5.2). There were no significant differences in plant height (F4, 15 = 1.38, P = 0.29) and number of leaf twig layers (F4, 15 = 2.78, P = 0.07) on plants prior to simulated herbivory damage. Ant recruitment on leaves Both cutting and punching of young leaves resulted in a 2–4‐fold increase in the number of ants compared to undamaged controls (Fig. 5.2; repeated measures ANOVA: F4, 95=15.10, P < 0.001). In both treatments, ant recruitment varied significantly over time (Fig. 5.2; repeated measures ANOVA: F2.95, 279.79=84.76, P < 0.001) and reached a peak 6 h after damage and remained elevated until the end of the 24‐h observation period. Interactions between treatment and time were highly significant for both types of treatment. Cutting and punching damage of young leaves had similar effects (P > 0.05). The two damage treatments of old leaves did not result in any significant changes in ant numbers compared to controls (Fig. 5.2; Dunnett’s contrast, P>0.05). There were significantly more active ants in the cutting treatment (F1, 38 = 15.027, P < 0.001) as well as in the punching treatment (F1, 38 = 23.521, P < 0.001) of young leaves than old leaves. 93
Chapter 5 The role of EFNs & ants on Leea 30
n=20
CY
CO
PY
PO
CK
Number of ants/leaf
25
20
15
10
5
0
6h
10min
Time after treatment
0
24h
Fig. 5.2 Mean numbers of ants (±SE) on young and old L. glabra leaves damaged by cutting and punching at different times after treatment compared to undamaged control leaves. t0 corresponds to the numbers immediately before damage. 30
n=20
Paper discs
Leaf discs
Number of ants/leaf
25
20
15
10
5
0
0
6h
10min
Time after treatment
24h
Fig. 5.3 Mean numbers of ants (±SE) on young L. glabra leaves with fixed disks from young leaves of another L. glabra plant and with fixed paper disks. t0 corresponds to the numbers immediately before treatments. Disks of fresh young L. glabra leaves that were pinned on young leaves of another plant led to a significant increase in the number of ants compared to controls with fixed paper disks (Fig. 5.3; repeated measures ANOVA: F1, 38 = 16.306, P < 0.001).
94 Chapter 5 The role of EFNs & ants on Leea Tannin concentrations and herbivory The tannin concentration of young leaves was significantly higher than that of both medium‐aged and old leaves (MANOVA: F2, 33 = 67.074, P <0.001) exceeding 100% in both comparisons (Fig. 5.4a). The portion of damaged leaf area of young leaves was significantly lower than that of both medium‐aged and old leaves (Fig.5.4b; MANOVA: F2, 33 = 45.890, P < 0.001). 20
20
a
a
a
b
18
a
16
14
12
10
b
8
b
6
4
2
0
Damaged portion of leaf (%)
Leaf tannin concentration (%)
18
16
14
12
10
8
6
4
b
2
Young
Medium
Leaf age
Old
0
Young
Medium
Old
Leaf age
Fig. 4 Total leaf tannin concentration (a) and damaged leaf area ratio (b) in leaves of different age (percent; mean+SE). Columns marked with different letters are significantly different at P < 0.001 (MANOVA). 5.4 Discussion Induced changes of extrafloral nectar quantity or quality following natural or artificial tissue damage have been reported from a number of EFNs plants (Heil et al. 2000, 2001; Ness 2003; Wäckers et al. 2001; Wäckers & Bonifay 2004; Wäckers & Wunderlin 1999), but there are also plants that fail to show such induction (Koptur 1989; Smith et al. 1990). Similarly, in our experiments with L. glabra, we found no temporal changes in extrafloral nectar volume or nectar sugar concentration within 60 h after artificial damage, indicating there was no induction following the treatments. The nectar production differences between plants with cut young leaves can most probably be attributed to biotic or abiotic factors. Environmental conditions including the air humidity and temperature in the experiment site are the most probable factors which influenced the nectar quality and quantity measured. We can exclude the possibility that ants consumed nectar, though it was not possible to completely prevent the access of ants to the experimental plants in the field. In total, we found no indication that plants react to the damage of leaves by changes in 95
Chapter 5 The role of EFNs & ants on Leea extrafloral nectar quantity or quality. We acknowledge that the power of our tests is relatively low and results should be considered within this context. Though it did not completely remove the relatively low power of the test with a sample size of four plants each treatment in this measurement, there were no significant differences of main effect within the tests. Despite this lack of nectar changes, our field experiments showed a significant increase in ant abundance on plants with damaged young leaves within 24 h after damage, although not on plants with damaged old leaves. This indicates that herbivore‐specific elicitors are not required for induction in Leea‐ants system. These results are consistent with those of Heil et al. (2001) who found induction of plants by artificial damage. They differ slightly from those of the Amazonian ant–plant Hirtella myrmecophila which showed that leaf wounds induced ant recruitment regardless of the leaf’s age (Romero & Izzo 2004). In our experiments, we did not observe a direct link between EFN production and ant abundance on damaged leaves. According to field observations, the ants attracted to damaged plants directly moved to the damaged young leaves and ignored the extrafloral nectaries at the leaf petioles. Extrafloral nectar production of L. glabra therefore seems to be a constitutive defense mechanism by the plant, suitable to create a general attractiveness to ants by the consistent production of extrafloral nectar provided as food. This interpretation is supported by our observation that extrafloral nectaries are only active at young leaves, which are most resistant to herbivore damage. As suggested by the results of the leaf disk experiment, the increased recruitment of ants is most probably caused by the release of VOCs by damaged young leaves of L. glabra, indicating an inducible mechanism of herbivore defense. A number of other studies also clearly indicate that ants are able to detect VOCs of damaged leaves and can distinguish between damaged plant species (Agrawal 1998b; Bruna et al. 2004; Inui & Itioka 2007). Furthermore, the tannin concentration of young leaves is more than 100% higher than in medium‐aged and old leaves, and the portion of damaged leaf area of young leaves is lower than for both groups of medium‐aged and old leaves; this maybe indicates the existence of an additional constitutive protective mechanism of the plant. There is evidence that tissues with high tannin content have lower numbers of herbivorous insects and lower damage levels than tissues with lower tannin content (Bialczyk 1999; Coley 1986). 96 Chapter 5 The role of EFNs & ants on Leea Our results support the general assumption of the optimal defense hypothesis that the most valuable tissues of a plant receive a higher proportion of the overall defensive investment than tissues of lower value. Young leaves of L. glabra which suffer low‐level damage by herbivores are more defended than old leaves. This looks contradict to some other tropical trees which suffer more damage by herbivores on young leaves than on old leaves (Kursar & Coley 2003). Maybe it was attributed to the plant tissues using divergent defensive strategies such as “escape” as well as “defense.” Many plants have multiple defense mechanisms for adaptation to different selection pressures (Coley et al. 2005) and there are not always trade‐offs among them (Koricheva et al. 2004). Our results show that young leaves of L. glabra are protected against attacks by herbivores by different mechanisms which include: (1) the constitutive activity of EFNs during the period of young leaf production which attract ants from the surrounding ground; (2) a mechanism induced by the damage of young leaves which leads to increased ant recruitment, most probably caused by the release of VOCs and (3) a higher allocation to tannins in young leaves than in older leaves. Young leaves of the L. glabra plants have multiple defense mechanisms, which indicate that a these young tissues are of significant value to the plant. Further detailed studies are still needed to fully understand the production of defenses in L. glabra, including experiments to reveal the factors influencing EFN production and leaf tannin and anthocyanin concentration such as the effects of different nutrient and light conditions. Many plants have multiple defense mechanisms for adaptation to different nutrient and light conditions. 5.5 References Agrawal AA (1998a). Leaf damage and associated cues induce aggressive ant recruitment in a neotropical ant‐plant. Ecology, 79, 2100‐2112. Agrawal AA (1998b). Induced responses to herbivory and increased plant performance. Science, 279, 1201‐1202. Agrawal AA, Dubin‐Thaler BJ (1999). Induced responses to herbivory in the Neotropical ant‐plant association between Azteca ants and Cecropia trees: response of ants to potential inducing cues. Behavioral Ecology and Sociobiology, 45, 47‐54. Agrawal AA, Rutter, MT (1998). Dynamic anti‐herbivore defense in ant‐plants: the role of 97
Chapter 5 The role of EFNs & ants on Leea induced responses. Oikos, 83, 227‐236. Bialczyk J, Lechowski Z, Libik A (1999). The protective action of tannins against glasshouse whitefly in tomato seedlings. Journal of Agricultural Science, 133, 197‐201. Bruna EM, Lapola DM, Vasconcelos HL (2004). Interspecific variation in the defensive responses of obligate plant‐ants: experimental tests and consequences for herbivory. Oecologia, 138, 558‐565. Chamberlain SA, Holland JN (2009). Body size predicts degree in ant‐plant mutualistic networks. Functional Ecology, 23, 196‐202. Coley PD (1986). Costs and benefits of defense by tannins in a neotropical tree. Oecologia, 70, 238‐241. Coley PD, Lokvam J, Rudolph K, Bromberg K, Sackett TE, Wright L, Brenes‐Arguedas T, Dvorett D, Ring S, Clark A, Baptiste C, Pennington RT, Kursar TA (2005). Divergent defensive strategies of young leaves in two species of inga. Ecology, 86, 2633‐2643. Grangier J, Dejean A, Male PJG., Orivel J (2008). Indirect defense in a highly specific ant‐plant mutualism. Naturwissenschaften, 95, 909‐916. Heil M (2008). Indirect defence via tritrophic interactions. New Phytologist, 178, 41‐61. Heil M, Fiala B, Baumann B, Linsenmair KE (2000). Temporal, spatial and biotic variations in extrafloral nectar secretion by Macaranga tanarius. Functional Ecology, 14, 749‐757. Heil M, Koch T, Hilpert A, Fiala B, Boland W, Linsenmair KE (2001). Extrafloral nectar production of the ant‐associated plant, Macaranga tanarius, is an induced, indirect, defensive response elicited by jasmonic acid. Proceedings of the National Academy of Sciences of the United States of America, 98, 1083‐1088. Inui Y, Itioka T (2007). Species‐specific leaf volatile compounds of obligate Macaranga myrmecophytes and host‐specific aggressiveness of symbiotic Crematogaster ants. Journal of Chemical Ecology, 33, 2054‐2063. Koptur S (1989). Is extra‐floral nectar production an inducible defense? In: Bock JH, Linhart YB (ed) The evolutionary ecology of plants. Westview, Boulder, Colo, pp 323‐339. Koptur S (1992). Interactions between insects and plants mediated by extrafloral nectaries. In: Bernays EA (ed) CRC Series on Insect/Plant Interactions, Vol. 4. CRC Press, Boca Raton, FL, pp 85‐132. Koricheva J, Nykanen H, Gianoli E (2004). Meta‐analysis of trade‐offs among plant antiherbivore defenses: Are plants jacks‐of‐all‐trades, masters of all? American Naturalist, 163, E64‐E75. Kursar TA, Coley PD (2003) Convergence in defense syndromes of young leaves in tropical 98 Chapter 5 The role of EFNs & ants on Leea rainforests. Biochemical Systematics and Ecology, 31, 929‐949. Li CL (1996). New taxa in Vitaceae from China. Chinese Journal of Applied Environmental Biology, 2, 43‐53. Mckey D (1979). The distribution of secondary compounds within plants. In: Rosenthal G.A, Janzen DH (ed) Herbivores: Their Interaction with Secondary Plant Metabolites. Academic Press, Orlando, FL, pp 56‐134. Ness JH (2003). Catalpa bignonioides alters extrafloral nectar production after herbivory and attracts ant bodyguards. Oecologia, 134, 210‐218. Pulice CE, Packer AA (2008). Simulated herbivory induces extrafloral nectary production in Prunus avium. Functional Ecology, 22, 801‐807. Raupp MJ, Denno RF (1983). Leaf age as a predictor of herbivore distribution and abundance. In: Denno RF, Mcclure MS (ed) Variable plants and herbivores in natural and managed systems. Academic Press, New York, pp 91‐124. Rhoades DF (1979). Evolution of plant chemical defense against herbivores. In: Rosenthal G.A, Janzen DH (ed) Herbivores: Their Interaction with Secondary Plant Metabolites. Academic Press, Orlando, FL, pp 4–55. Romero GQ, Izzo TJ (2004). Leaf damage induces ant recruitment in the Amazonian ant‐plant Hirtella myrmecophila. Journal of Tropical Ecology, 20, 675‐682. Rosumek FB, Silveira FAO, Neves FD, Barbosa NPD, Diniz L, Oki Y, Pezzini F, Fernandes G.W, Cornelissen T (2009). Ants on plants: a meta‐analysis of the role of ants as plant biotic defenses. Oecologia, 160, 537‐549. Smith LL, Lanza J, Smith G.C (1990). Amino‐acid‐concentrations in extrafloral nectar of impatiens‐sultani increase after simulated herbivory. Ecology, 71,107‐115. Wäckers FL, Bonifay C (2004). How to be sweet? Extrafloral nectar allocation by Gossypium hirsutum fits optimal defense theory predictions. Ecology, 85, 1512‐1518. Wäckers FL, Wunderlin R (1999). Induction of cotton extrafloral nectar production in response to herbivory does not require a herbivore‐specific elicitor. Entomologia Experimentalis Et Applicata, 91, 149‐154 Wäckers FL, Zuber D, Wunderlin R, Keller F (2001). The effect of herbivory on temporal and spatial dynamics of foliar nectar production in cotton and castor. Annals of Botany, 87, 365‐370. Zar JH (1996) Biostatistical analysis, 3rd edn. Prentice Hall, Upper Saddle River, NJ 99
Chapter 5 The role of EFNs & ants on Leea 100 Chapter 6 General Discussion Chapter 6 General Discussion 101
Chapter 6 General Discussion 6
General Discussion: Effects of habitat fragmentation and land use change on insect diversity in tropical landscapes The overall objective of the present study was to contribute to the knowledge on species interactions and functional diversity in a fragmented tropical landscape of southern Yunnan, China, at different scales. Insects have close direct and indirect relationships with plants, especially in tropical forests. The high plant species diversity of tropical forests leads directly to a higher diversity of leaf‐eating insects, and host specificities and plant traits can affect the insect preference (host plant selection, oviposition, feeding behavior) or performance (growth rate, development, reproductive success; Novotny et al., 2006, Novotny et al., 2007). The diverse host specificity in tropical rainforest also produces complex interactive relationships between the plants and insects. Therefore, insects are highly susceptible to adverse effects of forest fragmentation (Arnold & Asquith, 2002), and there is also evidence that tropical forest fragmentation reduces insect herbivory (Ruiz‐Guerra et al., 2010). Furthermore, forest fragmentation does not only decrease insect diversity directly, but truncate the food chains of specialized species (Komonen et al., 2000), and affect the trophic processes of highly complex food webs of the organisms involved (Dupont & Nielsen, 2006, Valladares et al., 2006). 6.1 General effects of habitat fragmentation on species diversity and abundance Adverse effects of habitat fragmentation on species diversity and abundance are basically explained by the reduction of the original forest habitat, resulting in fragmented forest patches of different size and distance or isolation between each other. More specifically, however, important research questions in studies on habitat fragmentation and landscape change refer to responses of habitat fragmentation of individual species populations and to the analysis of the factors affecting the population changes. Habitat fragmentation per se is a landscape‐level phenomenon in which species that survive in habitat remnants are confronted with a modified environment of reduced area, increased isolation and novel ecological boundaries. The implications of this for individual organisms are many and varied, because species with differing life history strategies are differentially affected (Ewers & Didham 2006). It is also assumed that 102 Chapter 6 General Discussion the effects of habitat fragmentation and ecological stress may be more pronounced in tropical systems than in temperate systems and may result in a greater proportional loss of local biodiversity in the former (Basset, 1996). Habitat fragmentation implies four effects of landscape change: (a) reduction in habitat area, (b) increase in number of habitat patches, (c) decrease in sizes of habitat patches, and (d) increase in isolation of patches, causing the disruption of species distribution patterns and forcing dispersing individuals to traverse a matrix habitat that separates suitable habitat fragments from each other (Fahrig, 2003). In addition, two other factors may affect species distribution and mobility in fragmented landscapes, which are the edge effect and the matrix effect. Habitat edges often alter the structure and diversity of invertebrate communities. Typically, species richness is negatively correlated with distance from the fragment edge into the fragment interior. The most common explanation for this trend is that there is a mixing of distinct fragment and matrix faunas at habitat edges, giving rise to a zone of overlap with greater overall species richness. Habitat edges can alter the nature of species interactions and thereby modify ecological processes and dynamics such as herbivory, seed predation and competition at a wide range of scales (Ewers & Didham 2006). The matrix effect refers to the structure and the available resources of the habitat types established between the original habitat patches. Depending on its nature, the matrix can be alternative or secondary habitat and conduct or hinder dispersal Matrix type thus may control the nature and magnitude of edge, area, and isolation effects and may regulate the use of corridors and stepping stones (Prevedello & Vieira 2010). The matrix habitat is a strong determinant of fragmentation effects within remnants because of its role in regulating dispersal and dispersal‐related mortality, the provision of spatial subsidies and the potential mediation of edge‐related gradients (Ewers & Didham 2006). Prevedello and Vieira (2010) analyzed 104 studies that compared effects of different matrix types on individuals, species and communities, covering broad range of landscape types and spatial scales. The type of matrix surrounding habitat patches affected species abundance or diversity in 95% of the studies, but such effects were overall smaller compared to patch size or isolation effects. Overall, Prevedello and Vieira (2010) found that the type of matrix is important, but patch size and isolation are the main determinants of ecological parameters in landscapes. Matrix quality generally increases with increasing structural similarity with habitat patches, a pattern that could be used as a general 103
Chapter 6 General Discussion guideline for management of the matrix in fragmented landscapes. The area and connectivity of habitat fragments is most important for the conservation of habitat specialists, whereas generalists may profit from a diverse surrounding landscape matrix (Steffan‐Dewenter, 2003). The long‐term effects of fragmentation are relatively poorly known as most studies of anthropogenically fragmented landscapes have been conducted less than 100 years after fragmentation While some authors consider time‐scales of 50 to 90 years as ‘long‐term’ and sufficient to ensure that diversity patterns have reached a dynamic equilibrium, this time frame may not be long enough to allow the full spectrum of fragmentation effects to be exhibited (Ewers & Didham 2006). Overall, the effects of landscape fragmentation are still poory understood. More studies of the independent effects of habitat loss and fragmentation per se are needed to determine the factors that lead to species, population and community effects (Fahrig 2003, Tschantke & Brandl 2004). Ewers and Didham (2006) emphasise that anthropogenic fragmentation is a recent phenomenon in evolutionary time and suggest that the final, long‐term impacts of habitat fragmentation may not yet have shown themselves. 6.2 Effects of habitat fragmentation and land use change on species diversity and abundance in the study area The study area of the Naban River valley in Xishuangbanna, southern Yunnan, was originally covered by tropical rain forest. Traditional agriculture, practiced since many decades, includes rice production in the lowlands and slash‐and‐burn farming on mountainous slopes. Land use was characterized by interactions of mixed systems, in which forests were maintained. Within the last decade, continued expansion of rubber cultivation took place and now most of the valley area is covered by rubber plantations. However, rubber plantations do not represent a uniform type of land use, but rather a spatio‐temporal dynamic system, ranging from young and open to closed canopy stands of very different ecological conditions and plant species. Stands of different age exist at the same time within a rotation cycle of about 40 years. The remaining land use types in the valley include secondary and primary forest fragments, grassland and shrubland successions as well as rice fields in the valley bottom along the river. Overall, the landscape presently represents a highly dynamic 104 Chapter 6 General Discussion situation, with increasing loss and fragmentation of forests and rapidly changing matrix patterns of closed and open habitats, largely with continuously changing spatial and temporal boundaries between habitat and land use types. 6.2.1 Effects on carabid beetles The overall result on the carabid beetle distribution patterns of the valley landscape indicates that three habitats types (rice field fallows, young open successions and natural forest) possess a degree of uniqueness in species composition, each characterized by species with significant indicator values. Since the natural vegetation in the study region is tropical forest, it can be concluded that the specialist and certainly more species from the arable land originate from naturally open habitat types and probably colonized the valley of the study area along rivers of the upper Mekong catchment area from the north. Their original habitats may be represented by riverbanks and natural grasslands, from where they followed the human land cultivation similar to the process in Europe. This can also be assumed for the ground beetle species typical of the grassland and shrubland in the study area. Carabid communities of young rubber plantations were quantitatively similar to those of forests, but without species of significant indicator value. With increasing plantation age, the number of carabid species decreased. The results indicate that the increasing age and a further spatial expansion of rubber plantations at the expense of forest areas will have negative impacts on the native forest carabid communities with strongest effects on forest specialists and rare species. 6.2.2 Effects on pollinators (wild bees and hoverflies) Wild bees and hoverflies also showed significant differences in the response to habitat loss and landscape fragmentation. Hoverflies were most common in young successional stages with highest number in the rice field fallows and rare in forests, and species richness was closely correlated with the number of flowering forb species in the different habitat types. This relationship can be explained by the resources provided by the plants, i.e., nectar and pollen serving as food for the adults of all hoverfly species. Flowering forb species can be considered as indicators for the “openness” of a habitat, as they usually decreased with proceeding succession from agricultural land to forest as well as from young to old rubber plantations with increasing canopy coverage. This was confirmed by the result that species number 105
Chapter 6 General Discussion and abundance of hoverflies are higher in young rubber plantations (5 and 8 years) than in older ones (20 and 40 years) and in forests. The natural tropical forest was not the habitat source of the vast majority of the hoverflies recorded in this area. Rather, its reduction by land cultivation favored the richness and distribution of hoverflies, indicating that most of the species collected might not originate from forest sites. In contrast to the hoverflies, wild bees showed significantly higher numbers of species and individuals in forests than in any other types of habitat. The latter had high similarities in bee community composition among each other but no close correlation with flowering forbs and other environmental variables. Although most wild bee species were recorded from at least one of the forest sites, most of these species were additionally recorded from other habitat types. This indicated that many bee species showed low specificities for habitat and floral resources and high abilities to move within the landscape. Although floral resources are of general importance, their relationship to bees can be influenced by landscape patterns and disturbance effects. In addition, wild bee communities are composed of species with different habitat and resource requirements, due to differences in the use of flowering plants, nesting requirements, dispersal modes, and other traits of species related to landscape structure A further expansion of rubber cultivation will result in large areas of mature rubber plantations. Because rubber cultivation largely proceeded at the expense of forest areas and not agricultural land, it can be assumed that hoverfly communities will not be negatively affected by this development. However, about a quarter of wild bee species were only recorded from forests, indicating that natural forest habitats were necessary to sustain the populations of many wild bee species. The large number of wild bee species which were recorded from forest can be negatively affected in a landscape with increasing rubber plantations. 6.2.3 Effects on species interactions Results of the study on animal‐seed interactions of a wild banana species (Musa acuminata) in forest and shrubland habitats showed differences in the diversity and abundance of seed dispersing and seed consuming species referring to temporal (seasons) and spatial (site and habitat) factors. To estimate the proportion of seed removed by secondary seed dispersers (ants) and seed predators (rodents), seeds 106 Chapter 6 General Discussion were artificially exposed in forest sites with and without M. acuminata stands and in open land habitats. Overall, primary seed dispersal showed a high dependency from frugivore volant animals (mainly bats) and can be considered as the most relevant factor in the fate of the M. acuminata seeds in the study region. However, the largest proportion of primarily dispersed seeds is consumed by rodents, but rodents also contribute to a decrease in competition between parent plants and seedlings. In addition to the primary dispersers, ants are shown to be important secondary seed dispersers in the open land habitats, contributing to the escape of seeds from post dispersal predation. The highest seed predation rate by rodents (70%) was found in forest with M. acuminata stands, corresponding with the highest rodent diversity (species numbers and abundance) among the habitat types. In contrast, the seed removal rate by ants was highest in the open land habitats, but there was no close correlation with ant diversity. Seed removal rates by ants were significantly higher in the dry compared to the rainy season, but rodent activity showed no differences between seasons. The overall results suggest that the largest proportion of seeds produced by M. acuminata are primarily dispersed by volant animals (bats), but most of the dispersed seeds were then consumed by rodents. 6.3 References Arnold AE, Asquith NM (2002). Herbivory in a fragmented tropical forest: patterns from islands at Lago Gatun, Panama. Biodiversity and Conservation, 11, 1663‐1680. Basset Y (1996). Local communities of arboreal herbivores in Papua New Guinea: Predictors of insect variables. Ecology, 77, 1906‐1919. Dupont YL, Nielsen BO (2006). Species composition, feeding specificity and larval trophic level of flower‐visiting insects in fragmented versus continuous heathlands in Denmark. Biological Conservation, 131, 475‐485. Ewers RM, Didham RK (2006). Confounding factors in the detection of species responses to habitat fragmentation. Biological Reviews of the Cambridge Philosophical Society, 81, 117‐142. Fahrig L (2003). Effects of habitat fragmentation on biodiversity. Annual Review of Ecology Evolution and Systematics, 34, 487‐515. Novotny V, Drozd P, Miller SE, Kulfan M, Janda M, Basset Y, Weiblen GD (2006). Why are there so many species of herbivorous insects in tropical rainforests? Science, 313, 1115‐1118. Novotny V, Miller SE, Hulcr J, Drew RAI, Basset Y, Janda M, Setliff GP, Darrow K, Stewart AJA, 107
Chapter 6 General Discussion Auga J, Isua B, Molem K, Manumbor M, Tamtiai E, Mogia M, Weiblen GD (2007). Low beta diversity of herbivorous insects in tropical forests. Nature, 448, 692‐698. Prevedello JA, Vieira MV (2010). Does the type of matrix matter? A quantitative review of the evidence. Biodiversity and Conservation, 19, 1205‐1223. Ruiz‐Guerra B, Guevara R, Mariano NA, Dirzo R (2010). Insect herbivory declines with forest fragmentation and covaries with plant regeneration mode: evidence from a Mexican tropical rain forest. Oikos, 119, 317‐325. Steffan‐Dewenter I (2003). Importance of habitat area and landscape context for species richness of bees and wasps in fragmented orchard meadows. Conservation Biology, 17, 1036‐1044. Strong DR, Lawton JH, Southwood TRE (1984). Insects on Plants: Community Patterns and Mechanisms. Cambridge, MA: Harvard University Press. Tscharntke T, Brandl R (2004). Plant–insect interactions in fragmented landscapes. Annual Review of Entomology, 49, 405‐430. 108 Summary 7
Summary: Animal‐plant‐interactions at different scales in changing tropical landscapes of southern Yunnan, China Southeast Asia is experiencing the highest relative rates of deforestation and forest degradation in the humid tropics due to logging and agricultural expansion. In particular, growing global demands for renewable products and commodities are driving the rapid expansion of large monocultures including rubber plantations at the expense of natural forests. These anthropogenic impacts are expected to result in species diversity declines and species extinctions as well as in disruptions of functional diversity. To increase the understanding of the effects of forest habitat degradation and fragmentation on those attributes on different scales, different studies were conducted in fragmented tropical landscapes of southern Yunnan, China, where large areas of forest have been, and still are, successively transformed into commercial rubber monoculture plantations. Specific objectives were to analyze (1) the species richness of ground beetle (Carabidae) communities within a landscape mosaic, (2) diversity of insect pollinators (wild bees and hoverflies) in relation to habitat type and flowering resources, (3) primary and secondary seed dispersal and seed predation in the wild banana species Musa acuminata and (4) mutualistic interactions between the forest understorey shrub Leea glabra and ants. Study 1 was conducted at 13 sites of different types of habitat (rice field fallows, early natural successions, rubber plantations, natural forest) over different seasons. In total, 102 species of Carabidae (including Cicindelinae) were recorded. Environmental factors explaining 80% of the total variation in carabid assemblage composition are the degree of vegetational openness of a habitat and its plant species diversity. Rice field fallows had highest numbers of species and individuals and are dominated by species probably originating from other regions. With increasing age of rubber plantations, carabid species richness decreased. It is concluded that a further expansion of rubber plantations at the expense of forest areas will have negative impacts on the native forest carabid assemblages with strongest effects on forest specialists and rare species. Study 2 was conducted in the same area and showed that hoverflies (total 53 species) were most common in young successional stages of vegetation including rice field 109
Summary fallows and shrubland. Species richness was highest in rice field fallows and lowest in forests and showed a highly significant relationship with the number of forb species and ground vegetation cover. In contrast, the highest richness of wild bees (total 44 species) was recorded from the natural forest sites, which showed a discrete bee community composition compared to the remaining habitat types. There was no significant relationship between the bee species richness and the environmental variables including the numbers of different plant life forms, vegetation cover, successional stage or land use type. At landscape scale, open land use systems including young rubber plantations are expected to increase the species richness of hoverflies, but result in negative effects on wild bee species diversity. Study 3 was conducted to estimate the proportion of Musa acuminata seeds removed by primary and secondary seed dispersers and seed predators. Seeds were experimentally exposed in forest sites with and without M. acuminata stands and in open land habitats. Overall, primary seed dispersal showed a high dependency from frugivore volant animals (mainly bats). However, the largest proportion of primarily dispersed seeds is consumed by rodents, but rodents also contribute to a decrease in competition between parent plants and seedlings. In addition to the primary dispersers, ants are shown to be important secondary seed dispersers in the open land habitats, contributing to the escape of seeds from post dispersal predation. The highest seed predation rate by rodents (70%) was found in forest with M. acuminata stands, corresponding with the highest rodent diversity (species numbers and abundance) among the habitat types. In contrast, the seed removal rate by ants was highest in the open land habitats, but there was no close correlation with ant diversity. Seed removal rates by ants were significantly higher in the dry compared to the rainy season. Study 4 included field experiments on Leea glabra in its natural forest habitat to analyze the effects of artificial damage of young and old leaves on extrafloral nectaries (EFNs) secretion quantity and sugar concentration, as well as the effects on ant abundance on the plants following the damage treatments. There were no rapid changes in extrafloral nectar volume or nectar sugar concentration which would indicate an induced reaction following artificial damage. However, both cutting and punching of young leaves resulted in a significant increase (2–4‐fold) of ants within 6 h after damage compared to undamaged controls. Another experiment indicated that ants were most probably attracted by volatile organic compounds (VOCs) released from damaged young leaves. Furthermore the proportion of damaged leaf area of 110 Summary young leaves was significantly lower than that of old leaves and the concentration of tannins was significantly higher in young than in medium and old leaves. In conclusion, results show that young leaves of L. glabra are protected against attacks by herbivores by multiple mechanisms, which include: (1) the activity of EFNs, which attract different ant species from the surrounding ground; (2) a mechanism induced by the damage of young leaves, which leads to rapidly increased ant recruitment and (3) a higher allocation of tannins in young than in older leaves. 111
Summary 112 Zusammenfassung 8 Zusammenfassung: Tier‐Pflanze‐Interaktionen auf verschiedenen Skalen tropischer Landschaften Süd‐Yunnans (China) im Wandel Südostasien ist von den höchsten relativen Raten an Waldverlust und Walddegradation in den feuchten Tropen betroffen, bedingt durch Abholzungen und die Ausdehnung landwirtschaftlicher Nutzflächen. Insbesondere geht die global steigende Nachfrage nach erneuerbaren Energien und Rohstoffen, die zu einer starken Ausdehnung großer Monokulturen wie Kautschukplantagen führt, auf Kosten natürlicher Wälder. Diese anthropogenen Veränderungen können auch zu einem Rückgang der Artendiversität und dem Aussterben von Arten sowie zu negativen Veränderungen in der funktionellen Diversität führen. Um die Auswirkungen der Degradation und Fragmentierung von Waldhabitaten auf diese Merkmale auf verschiedenen Skalenebenen besser zu verstehen, wurden verschiedene Untersuchungen in fragmentierten tropischen Landschaften Süd‐Yunnans (China) durchgeführt, wo große Waldflächen fortschreitend durch kommerzielle Kautschuk‐Monokulturen ersetzt werden. Die Ziele im Einzelnen waren die Analyse (1) der Artenvielfalt von Laufkäfer‐Gemeinschaften (Carabidae) innerhalb des Landschaftsmosaiks, (2) der Diversität bestäubender Insektengruppen (Wildbienen und Schwebfliegen) in Abhängigkeit vom Habitattyp und den Blütenressourcen, (3) der primären und sekundären Samenverbreitung und der Samenprädation bei der Wildbananenart Musa acuminata und (4) der mutualistischen Interaktionen zwischen dem Waldunterwuchsstrauch Leea glabra und Ameisen. Untersuchung 1 wurde an 13 Standorten verschiedener Habitattypen (Reisfeldbrachen, junge Sukzessionen, Kautschukplantagen, natürlicher Wald) über verschiedene Jahreszeiten durchgeführt. Insgesamt wurden 102 Carabidae‐Arten nachgewiesen (einschl. Cicindelinae). Von den Umweltfaktoren erklären der Grad der Vegetationsbedeckung und die Pflanzenartendiversität zusammen 80% der Variation in der Zusammensetzung der Carabidengemeinschaften. Reisfeldbrachen wiesen die höchsten Arten‐ und Individuenzahlen auf und werden von Arten dominiert, die wahrscheinlich aus anderen Regionen zugewandert sind. Mit zunehmendem Alter der Kautschukplantagen nahm die Artenvielfalt der Carabidae ab. Die Ergebnisse lassen den Schluss zu, dass eine weitere Ausdehnung der Kautschukplantagen auf Kosten von Waldflächen negative Auswirkungen auf die Carabidengemeinschaften der 113
Zusammenfassung Wälder ausübt, mit den größten Konsequenzen für Waldspezialisten und seltene Arten. Untersuchung 2 wurde in demselben Gebiet durchgeführt und zeigte, dass Schwebfliegen (insgesamt 53 Arten) ihre größte Häufigkeit in jungen Sukzessionsstadien einschließlich Reisfeldbrachen aufwiesen. Der Artenreichtum war am höchsten in Reisfeldbrachen und am geringsten in Wäldern und zeigte eine signifikante Beziehung zur Zahl der krautigen Blütenpflanzenarten und mit der Vegetationsbedeckung. Im Unterschied dazu wurde die höchste Vielfalt an Wildbienen (insgesamt 44 Arten) an den natürlichen Waldstandorten nachgewiesen, die eine eigenständige Artengemeinschaft gegenüber den anderen Habitattypen aufweisen. Signifikante Beziehungen zwischen bestimmten Umweltvariablen (Pflanzenzusammensetzung, Vegetationsbedeckung, Sukzessionsstadium oder Landnutzung) und der Artendiversität der Wildbienen wurden nicht gefunden. Auf Landschaftsebene fördern offene Standorte einschließlich junger Kautschukplantagen die Artenvielfalt der Schwebfliegen, wirken sich aber negativ auf die Wildbienendiversität aus. Untersuchung 3 wurde durchgeführt, um die Anteile der Samen von Musa acuminata‐Pflanzen zu bestimmen, die von primären und sekundären Samenverbreitern sowie von Samenkonsumenten entfernt werden. Samen wurden experimentell an Waldstandorten mit und ohne Bananenbestände sowie an Offenlandstandorten ausgelegt. Insgesamt zeigte die primäre Samenverbreitung eine hohe Abhängigkeit von flugfähigen Tieren (v.a. Flughunden). Der größte Anteil der primär verbreiteten Samen wird jedoch von Nagetieren konsumiert, die aber auch zu einer Verringerung der Konkurrenz zwischen Mutterpflanzen und Keimlingen beitragen. Zusätzlich zu den Primärverbreitern erwiesen sich Ameisen als wichtige Sekundärverbreiter an den Offenlandstandorten und tragen damit zur Verminderung der Samenprädation bei. Die höchste Samenprädationsrate durch Nager wurde in Wäldern mit M. acuminata‐Beständen festgestellt (70%) und geht einher mit der dort nachgewiesenen höchsten Dichte an Nagerpopulationen der verschiedenen Habitate. Der größte von Ameisen entfernte Anteil an Samen zeigte sich in den Offenlandstandorten, aber es bestand keine Korrelation mit der Diversität oder Abundanz von Ameisenarten. Der von Ameisen entfernte Anteil an Samen lag in der Trockenzeit signifikant höher als in der Regenzeit. Untersuchung 4 umfasste Feldexperimente an Leea glabra‐Sträuchern an ihren natürlichen 114 Waldstandorten zur Analyse der Auswirkungen künstlicher Zusammenfassung Beschädigungen von jungen und alten Blättern auf die Quantität der Sekretionen und die Zuckerkonzentrationen der extrafloralen Nektarien (EFNs) sowie auf die Abundanz von Ameisenauf den Pflanzen infolge der Beschädigungen. Es zeigten sich keine raschen Veränderungen im Nektarvolumen oder in der Zuckerkonzentration, die auf eine induzierte Reaktion der Pflanzen auf die Schädigungen hindeuten. Die Schädigungen hatten jedoch einen signifikanten Anstieg (2‐4‐fach) der Ameisendichte auf den Pflanzen innerhalb von 6 h zur Folge. Ein weiteres Experiment deutete darauf hin, dass die Ameisen höchstwahrscheinlich von flüchtigen organischen Substanzen (VOCs) angelockt wurden, die von den beschädigten Blättern freigesetzt wurden. Außerdem wurde nachgewiesen, dass der Anteil der natürlichen Blattschädigungen junger Blätter signifikant geringer war als an alten und mittelalten Blättern, und außerdem der Tanningehalt in jungen Blättern signifikant höher war als in den anderen Altersklassen. Schlussfolgernd zeigen die Ergebnisse, dass die jungen Blätter von L. glabra durch mehrere Mechanismen vor Angriffen durch Herbivoren geschützt sind, und zwar (1) die Sekretion von Nektar durch EFNs, die für Ameisen aus der Umgebung attraktiv sind, (2) einen Mechanismus, der durch die Beschädigung junger Blätter induziert wird und zu einer rascher Erhöhung der Zahl an Ameisen auf der Pflanze führt und (3) den höheren Gehalt an fraßhemmenden Tanninen in jungen gegenüber älteren Blättern. 115
Zusammenfassung 116 Acknowledgement Acknowledgements Many thanks should be given to my advisor apl Prof. Dr. Konrad Martin. If without his scientific guidance, kind support as well as his willingness to discuss relevant scientific questions at all times this thesis would not have been successfully completed. Furthermore, I would like to express my gratitude to Prof. Dr. Claus Zebitz and Prof. Dr. Martin Dieterich for their willingness to review my work. I need appreciate Prof. Dr. Jin Chen in Xishuangbanna Tropical Botanical Garden, Chinese Academic of Sciences (XTBG) a lot for his recommendation of my Ph. D study. His enthusiasm and keen thought for sciences influenced me a lot and let me enter the palace of natural sciences gradually. I would like to sincerely thank the “Living Landscape China” Project from the BMBF of Germany for the financial support that made this thesis possible. Therefore, great thanks go out to the people involved in this project but especially to the Prof. Dr. Joachim Sauerborn, Dr. Gerhard Langenberger, Ms Inga Haeuser and “Rose”, Ms Xuan Liu. They all were necessary prerequisites of my study work during the past three years. For the supply of measuring instruments used in the field, I would like to thank Prof. Dr. Qing‐Jun Li of XTBG for listening to my concerns and providing me with valuable advices while initiating the field experiments of extrafloral nectar observation. Moreover, a big thanks to Dr. Andreas Weigel and Dr. Frank Burger in Erfurt Natural History Museum of Germany, Dr. Mei‐Ying Lin and the vice director Dr. Jun Chen in Beijing Institute of Zoology, Chinese Academy of Sciences, Dr. Liang Tang in Shanghai Normal University and many other beetle specialists who helped to identify the insect species. Here, I would like to especially thank my classmate and friend Mr. Jing‐Xin Liu, who provided the precious vegetation data in NRWNNR. They all were solid foundation which made this thesis completed successfully. A great thanks goes out to all colleagues at the Institute of Plant Production and Agroecology in the Tropics and Subtropics (380b), but especially to the Dr. Reza Golbon, Dr. Anna Treydte, Dr. Marc Cotter, Elisabeth Zimmermann, Eva Schmidt, Franziska Harich and Paul Trumpf. You all let me enjoyed study life when I stayed in Hohenheim. Moreover, I want to express my gratitude to Dr. Matthias Hartmann, the director of Erfurt Natural History Museum of Germany for his warm welcome when I stay their for a short period. Additionally, I would like to thank all members of NRWNNR but especially to Mr. Yun Yang, Mr.Feng Liu, Mr.Guang‐Hong Cao, Mr. 117
Acknowledgement Zhi‐Ming Ma, Mr.Bing Ai, Mr.Guang Ai, Xi Ni, Zhou shifu and my field assistants, local Dai people Gong Ai and Kanjian Ai, who supported me during my stay within the natural reserval and helped me to do lots of hard field work. Dr. Yan Liu in Justus Liebig University of Giessen gave me another personal look to observe the world when we stay together in that small “LILAC” office at Jinghong. These acknowledgements would not be complete without a big thanks to my former colleagues Mr. Shou‐Hua Yin, Mr. Jia‐Yuan, Huang, Mr. Ming‐Ma, Ms.Jin‐Yan Kuang and Mr.Yu Song, I appreciate their warm welcome when I come back to XTBG every time. Also a big thanks should be given to my former supervisors, Prof. Zai‐Fu, Xu and Prof.Dr.Kun‐Fang, Cao in XTBG for their kind supports and confidences with me. Last but not least, I want to thanks my whole family in remote Anhui Province, especially Mum Zhu‐Yi Wu and Dad Xiao‐Qiao Meng for your continuous trust and unconditioned support where ever I was or what ever I did. Dad, your enthusiasm for physics science during my young time was the major reason that I determine to engage myself to research. Mum and Dad come to Xishuangbanna in 2009 and 2008 respectively that not only helped to take care of my daughter but also did a lot of housework for me when I stayed in Jinghong for field experiments and that made me can concentrate on my own study completely. Jian‐Zeng, my old brother, you are always there and keep eyesight on me. Your courage and optimism for tough life encouraged me a lot and gave me confidences to hold out the things what I enjoyed even in that the most difficult time without light. My wife, Ms. Jiao He, thanks for your accompanying with me and supports during that tough but glorious days. And finally, cheers to little Qian‐Yu, Yang‐Yang and Yang‐Yue. You three all are my sweat hearts and your smiles were on my mind forever. 118 Publication list Publication list Meng, L.‐Z., Martin K., Liu, J.‐X., Chen, J. (2011). Young leaf protection in the shrub Leea glabra in south‐west China: the role of extrafloral nectaries and ants. Arthropod‐Plant Interactions, 1‐7. DOI:10.1007/s11829‐011‐9151‐6. Meng, L.‐Z., Martin K., Weigel A., Liu, J.‐X., (2011). Impact of rubber plantation on carabid beetle communities and species distribution in a changing tropical landscape (southern Yunnan, China). Journal of Insect Conservation, DOI:10.1007/s10841‐011‐9428‐1. Meng, L.‐Z., Martin K., Liu, J.‐X., Burger F., Chen, J. (2012). Contrasting responses of hoverflies and wild bees to habitat structure and land use change in a tropical landscape (southern Yunnan, SW China). Insect Science, DOI: 10.1111/j.1744‐7917.2011.01481.x. Meng, L.‐Z., Gao, X.‐X., Chen, J., Martin K. (2012). Spatial and temporal effects on seed dispersal and seed predation of Musa acuminata in southern Yunnan, China. Integrative Zoology, DOI: 10.1111/j.1749‐4877.2011.00275.x. Meng, L.‐Z., K. Martin, A.Weigel, Liu, J.‐X., Chen, J., Lin, M.‐Y,(2012). The response of longhorn beetles (Coleoptera: Cerambycidae) to forest fragmentation in a changing tropical landscape (southern Yunnan, China), (in prep). Weigel A., Meng, L.‐Z., Lin, M.‐Y. (2011) Contribution to the Fauna of Longhorn Beetles (Coleoptera: Cerambycidae) in the Naban River Watershed National Nature Reserve (China: Yunnan, Xishuangbanna). (in Press).Book. Zhang, J.‐L., Meng, L.‐Z., Cao, K.‐F. (2009). Sustained diurnal photosynthetic depression in uppermost‐canopy leaves of four dipterocarp species in the rainy and dry seasons: does photorespiration play a role in photoprotection? Tree Physiology, 29(2):217–228. 119
Curriculum Vitae Curriculum Vitae Ling‐Zeng Meng born 17. 04. 1979 in Taihu Anhui, China May 2008‐Sep 2011 Ph.D candidate in Agroecology in the Tropics University of Hohenheim, Stuttgart, Germany under supervision of apl Prof. Dr. Konrad Martin in plant‐animal interactions Jul 2005‐Apr 2008 Research assistant in Xishuangbanna Tropical Botanical Garden, Chinese Academy of Sciences in Menglun Mengla Yunnan, China and major in plant‐animal supervision of Prof.Dr. Jin Chen Sep 2002‐Jun 2005 Master Student in in Xishuangbanna Tropical Botanical Garden, Chinese Academy of Sciences in Menglun Mengla Yunnan, China and major in plant physiology and conservation biology under supervision of Prof. Zai‐Fu Xu and Prof.Dr.Kun‐Fang Cao Sep 1998‐Jul 2002 Bachelor Student in North‐West Agriculture & Forestry University in Yangling Shanxi, China major in water and soil conservation 120 and Subtropics interactions (380b), under 1
2 Back Cover1 1
My daughter Qian‐Yu Meng (1y old) and father Xiao‐Qiao Meng (60y old) in Xishuangbanna, China. (May of 2008, Photograph by Ling‐Zeng Meng) Many years later, most probably I will be ashamed for my own ignorance and parochialness nowadays L‐Z Meng 2