• Asclepias lanceolata;
  • biodiversity;
  • community structure;
  • Everglades;
  • plant–insect interactions;
  • pollination;
  • pollinator;
  • reproductive success;
  • species richness;
  • tree islands


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information
  • 1
    The Everglades (Florida, USA) is a mosaic of different habitats. Tropical and temperate trees grow on patches of high ground (tree islands) surrounded by lower elevation wetland communities (marl prairie).
  • 2
    Tree islands of various sizes provide nesting substrate, larval host plants and floral resources for insect pollinators. Herbaceous plants in the open surrounding wetlands may also depend on these pollinators.
  • 3
    We investigated pollinator diversity and abundances in both tree island and marl prairie habitats using transect sampling methods and estimated pollination success of the milkweed Asclepias lanceolata, an insect-pollinated marl prairie species, in relation to distance from and size of the closest tree island.
  • 4
    On a total of 11 bayhead tree islands, we found that insect diversity and abundance were greater on the edge of larger tree islands (20–30 m2) than on smaller tree islands (5–10 m2). Pollinator diversity and abundance in the marl prairie decreased with increasing distance from tree islands.
  • 5
    Pairs of potted A. lanceolata plants were placed in the marl prairie at distances up to 1000 m from small and large tree islands. Fruit and seed production were highest for plants placed less than 25 m from tree islands and decreased with increasing distance.
  • 6
    Our results suggest that tree islands are an important source of pollinators for the plants in the tree island and surrounding wetland habitats.
  • 7
    This landscape-based study illustrates how overall landscape structure affects important biotic interactions, particularly plant–pollinator relationships. Our findings have far-reaching ecological implications for the reproductive success of plants in small, isolated populations that may depend on insect vectors for pollination.


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

The influence of landscape structure and habitat quality on insect behaviour, ecology and species interactions is fundamental to understanding the composition of insect communities (Schneider & Fry 2001; Tscharntke et al. 2002). Landscapes may comprise a mosaic of different habitats. Such habitat patches, especially when altered either naturally or anthropogenically, may be fragments that vary in size and quality. High-quality fragments are typically those that provide animals with suitable nesting sites, food resources, shelter and/or mating sites (Moertberg 2001; Perfecto & Vandermeer 2002). Differences in fragment density, size and quality may affect species-specific dispersal abilities within the landscape (Forman 1995; Ries & Debinski 2001), as well as the intensity of biotic interactions in the landscape, particularly seed dispersal (Masaki 2004; McEuen & Curran 2004), herbivory (Santos & Telleria 1994; Benitez-Malvido et al. 1999) and pollination (Aizen & Feinsinger 1994; Bronstein 1995). Both the degree of isolation of habitat fragments and their size may affect plant–animal interactions through changes in species diversity and abundance (Debinski et al. 2001; Summerville & Crist 2003).

Because many animals nest only in particular habitats, they may need to disperse into neighbouring habitats to breed and to find food. Accordingly, plants that depend on these animals for pollination or seed dispersal are also affected by the ability of pollen or seed vectors to move through the landscape (Tewksbury et al. 2002; Kreyer et al. 2004). As animals exploit multiple habitats to meet their various requirements, landscape-level approaches may provide insight into identifying how changes of physical environments at different spatial scales influence both plant and animal populations (Wiens 1997; Haynes & Cronin 2004).

The resources needed by many insect pollinators, including mates, nectar and pollen, nesting materials and larval food plants, may occur in different habitats in the landscape. For example, different pollinator guilds (bumblebees, honeybees and large wild bees) prefer patches of different qualities in the mosaic of habitats in an agricultural landscape (Hirsch et al. 2003). Strong fliers, such as bumblebees and honeybees, often forage great distances from their hives in order to exploit the most rewarding floral patches (Heinrich 1979; Visscher & Seeley 1982; Beekman & Ratnieks 2000). By contrast, weak-flying pollinators, such as solitary wild bees and some butterflies, may have limited ranges and forage for resources only near the nest or roosting site (Sutcliffe & Thomas 1996; Gathmann & Tscharntke 2002).

Heterogeneous landscapes and the spatial arrangement of habitats therein may also influence dispersal abilities and movement patterns of foraging insects that disperse into adjacent habitats. For example, different prairie edge types have been shown to influence the emigration rates of two butterfly species, a habitat generalist and a habitat specialist (Ries & Debinski 2001). Thus, the inherent nature of mosaic habitats often creates ecological barriers or movement corridors for many insect pollinators, and, as a result, impedes or facilitates dispersal ability to meet these species-specific requirements (Sutcliffe & Thomas 1996; Haddad 1999).

The ability of insect pollinators to move between habitats to fulfil their resource requirements may depend in part on their size, strength and flight ability. Studies of foraging ranges of insect pollinators in various habitats suggest that maximum flight distance varies greatly among species (Osborne et al. 1999; Beekman & Ratnieks 2000; Steffan-Dewenter & Kuhn 2003). For example, honeybees fly between 1 and 6 km away from the hive for food, depending on the floral richness of the habitat (Visscher & Seeley 1982; Waddington et al. 1994), while bumblebees fly up to 2.2 km for food (Kreyer et al. 2004). Although flight distance estimates for small-bodied, weak-flying pollinators are rare, the maximum foraging distances for several species of solitary bees range from 150 to 600 m, and foraging distance is positively correlated with length of the body (Gathmann & Tscharntke 2002). Therefore, pollinators that have long-distance flight capabilities, particularly large-bodied bees and large butterflies, may play a vital role in the reproduction of plants in the isolated areas of heterogeneous landscapes (Janzen 1971; Bawa 1990).

Plants that require pollinators to transfer pollen in isolated or fragmented habitats may be reproductively disadvantaged because pollen vectors may be in low density or have restricted flight capabilities (Ågren 1996; Steffan-Dewenter & Tscharntke 1999). In Asclepias tuberosa (Apocynaceae), annual differences in insect visitor abundances, together with differences in pollinator visitation rates, have been used to explain pollinator effectiveness (Fishbein & Venable 1996).

The Everglades landscape is a natural mosaic of mostly pristine habitats that are structured by the natural flow of water through the ecosystem (Davis & Ogden 1994; Lodge 1994). Tree islands, higher elevation outcrops bearing temperate and tropical shrubs and trees, pepper the landscape in various habitats and plant communities. This habitat type is particularly abundant in marl prairie, a marsh habitat characterized by a thin layer of calcitic soil (marl) covering the limestone substrate, which typically is inundated with water (hydroperiod) for 3–7 months. Marl prairie typically experiences the shortest hydroperiods of all marsh types in the Everglades (Olmsted et al. 1980; Lodge 1994).

We investigated the effects of the spatial arrangement and size of tree islands on the distribution, diversity and abundance of insect pollinators in both the tree island and the marl prairie communities. We also examined the effects of these patterns on the pollination success of Asclepias lanceolata, a wetland species growing at varying distances from tree islands of varying sizes. We selected A. lanceolata as our study species because it grows in marl prairie habitat but not in tree islands and we were able easily to quantify both removal of pollinia from flowers (Wyatt & Broyles 1994) and presence/absence of pollinia attached to appendages of foraging insects.

Materials and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

study area

Study sites were located in the Taylor Slough area of south-eastern Everglades National Park (ENP; 25°30′ N, 80°37′ W), Florida, USA. Data were collected between February and June 2002, during the dry season when winter temperatures ranged from 18 to 27 °C. Topographically low areas of Taylor Slough are characterized by sawgrass marshes and wet marl prairie communities, whereas higher elevations support islands of tropical and temperate trees and shrubs – tree islands (Kushlan 1990; Gunderson 1994).

Approximately 70–90% of the plant cover in marl prairie is Muhlenbergia filipes M. A. Curtis (sweetgrass), but more than 100 plant species, including the milkweed Asclepias lanceolata, are found in this community (Olmsted et al. 1980). The number of plant species is estimated at 4–13 m−2 (Olmsted et al. 1980). However, the total number of plant species increases with sampling area, suggesting that, with the exception of M. filipes, individual plant species are rather patchily distributed within marl prairies.

Tree islands vary in size, from small, slightly raised limestone mounds (< 1 m2) with only a few trees and/or shrubs to large areas of several hectares, composed of many species of trees, shrubs, vines and herbaceous plants (Davis & Ogden 1994; Sklar & van der Valk 2002). These islands also vary in species composition and their distances to other tree islands. The bayhead tree islands chosen for this study are typically composed of individuals of Persea borbonia L., P. palustris Raf. and Magnolia virginiana L. with some other trees and shrubs, such as Salix caroliniana Michx., Chrysobalanus icaco L., Ilex cassine L. and Myrica cerifera L., vines such as Sarcostemma clausum Jacq. and Mikania scandens L., and orchids, bromeliads and ferns (Davis & Ogden 1994; D. Artz, personal observation).

target study species

In the Everglades, A. lanceolata is not found in bayhead tree islands, but occurs naturally in marl prairie habitats where it flowers from early February through the summer months. Plants grow up to 1.5 m tall and typically have 1–3 terminal umbels with an average of seven flowers per umbel. Flowers range from yellow-orange to deep red and senesce after 5–7 days. Nectar production was not measured in the field, but we observed that the coronal hoods were usually nectar-rich; thus, A. lanceolata was probably a high-quality nectar resource for pollinators. In A. lanceolata, as in all milkweeds, pollen grains are stored as a collective unit called the pollinium, and one pollinium is sufficient to effect total seed set (see Wyatt & Broyles 1994 for milkweed pollination mechanism; Fishbein & Venable 1996).

tree island size and pollinator abundance and diversity

We selected five small (5–10 m2) and six large (20–30 m2) bayhead tree islands for transect sampling along the island edge. We did not sample in the interior of tree islands because pilot investigations suggested little pollinator activity there. All tree islands were approximately circular in shape and, within each size class, also displayed similar substrate elevation, vegetation composition and canopy structure. We visually counted or captured with a sweep net all potential insect pollinators seen within a 2.5-m-wide band on either side of the tree island perimeter transect and within 5 m in front of the circular transect around the island (see details of this method in Pollard 1977; Pollard & Yates 1993). We walked at a constant pace and recorded insect identity, time of day and plant species visited. To determine which insects are potential pollinators of A. lanceolata, we recorded the presence/absence of Asclepias pollinia. Any insect whose taxonomic identity could not be determined in flight was captured, and either identified and released or killed and brought back to the lab for identification: such insects were not recorded unless they were captured. Insect voucher specimens have been deposited in the South Florida Research Center museum, Everglades National Park.

Potential pollinators were sampled during 61 days between February and June 2002. Of the 11 tree islands under investigation, one small and one large tree island were randomly selected for sampling each day. Therefore, each tree island was sampled a total of 20–24 times through the season. The selected island was sampled once in the morning and once in the afternoon/early evening (around 08:00 and 19:00 h). Small tree islands (STIs) were sampled for 10 min each (i.e. 20 min per day). Large tree islands (LTIs) were sampled for 30 min, a total of 1 h per sampling day. We sampled on sunny days when the temperature exceeded 15 °C and winds were minimal.

Night-time sampling was not undertaken because pilot studies suggested that nocturnal moths do not visit A. lanceolata (D. Artz, unpublished data: mesh bags were placed over flowers prior to anthesis, but removed at dusk and flowers checked for pollinia removal/insertion early the following morning).

To explore differences in overall community structure of insect taxa among the tree islands, we calculated the accumulated insect species richness, total abundance, alpha diversity (H′), maximum species diversity (inline image) and evenness (J′) of tree islands after pooling for each particular size class. We calculated alpha diversity using the Shannon–Weiner diversity index, which is a measure that takes into account the proportional abundance of each species (Magurran 1988). To determine differences in pollinator community composition among tree islands of different sizes, we used Sorensen's index [S = (a + b)/2c, where a = species at one site, b = species at second site, c = species at both sites] (Krebs 1989).

marl prairies and their pollinator communities

In order to quantify pollinator species richness, composition and abundance in the marl prairie that surrounds tree islands, we sampled over 5 months (February–June) for 248 h using the ‘Pollard Walk’ transect sampling technique (Pollard 1977; Pollard & Yates 1993). Visual counts and sweep net collections were made along permanent transects that projected 1 km away from tree islands. During a sampling bout we walked away from a tree island and back along the same path (total 2 km) and recorded insect identification to the smallest possible taxonomic category, time of collection, behaviour (nectar/pollen foraging, ovipositing, resting), species of plant visited and distance away from the tree island. We sampled two transects at 1-h intervals for each tree island. Species richness was calculated as the cumulative total of all species, and abundance as the number of individuals recorded for all transects at each sampling site.

pollination success of a. lanceolata in the marl prairie

Plants used in this study resulted from seeds collected from three natural populations of A. lanceolata during August 2001 in Long Pine Key, an area south of the study area, and Chekika, a north-eastern section of ENP (25°37′ N, 80°35′ W). Seeds were sown in a shadehouse in soil collected from marl prairie habitat and seedlings transplanted into 1-gallon plastic pots. Prior to anthesis, the plants were assigned at random to experimental sites in the marl prairie; pots set on the ground were no more conspicuous than the surrounding natural vegetation. Herbivory was minimal throughout vegetative and reproductive phases.

To determine the effect of distance from tree islands on the pollination success, we established permanent sampling sites at four distances from tree islands: 25, 100, 300 and 700–1000 m along paths radiating in two directions from the same five small and six large tree islands as used to sample community composition. Two plants were placed 2 m apart at each sampling site. Where possible, we located the most distant site 1 km from the tree island but, because of the spatial arrangement of tree islands in the landscape, the maximum distance was 700 m away for one large tree island and 800 m for one small tree island.

We observed insect visits to each pair of A. lanceolata plants during 15-min periods from March to May. A visit occurred when an insect probed for nectar, oviposited or rested on flowers. All insects were captured after the visit for identification and inspection of the body for pollinia and then released. Taxonomically difficult species were collected and identified in the lab. We pooled the observational visitation data at each distance class for both small and large tree islands.

We investigated whether tree island pollinator diversity influenced visitation rates to A. lanceolata in the surrounding tree island – marl prairie matrix, and whether plants close to tree islands were visited by a larger suite of potential pollinators. We pooled the observational visitation data at each distance class for both small and large tree islands.

pollinator size and flight distance

To determine the relationship between larger-bodied, strong flying bees and visitation rates to A. lanceolata at greater distances away from tree islands, we measured the combined length of thorax and abdomen of at least five specimens for each bee species from the insect collections at the University of Miami (Coral Gables, Florida) and the South Florida Natural Resources Center (ENP) as an estimate of bee body size.

reproductive success of a. lanceolata

To determine pollination, together with several measures of reproductive success, as a function of distance from tree islands, we quantified fruit set per plant, seed set per plant and percentage germination of mature seeds. Seeds were collected by covering mature fruits with bridal veil mesh. We collected all developed follicles and counted the number of seeds per follicle and seeds per plant. Because paired plants were placed in close proximity to each other, pollinator visits to the experimental plants were probably not independent events; therefore, we averaged for each pair the resulting number of fruits and seeds per plant. The relationships between fruit production and percentage germination of seeds, and distance from tree islands, were analysed using one-way anova. Multiple comparison t-tests were used to determine differences in the effect of distance from tree islands. We used linear regression to determine the relationship between the number of seeds produced per plant and distance from the nearest tree island.


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

tree island size and pollinator abundance and diversity

Our analysis of alpha diversity shows that LTIs had higher insect species diversity and abundance than STIs. The total pollinator diversity differed significantly (H′ = 4.040 and 3.446, on LTIs and STIs, respectively; t982 = 41.24, P < 0.0001). Comparisons of the pollinator communities of STIs and LTIs suggested intermediate similarity (Sorensen's Index = 0.557). On STIs, 43 species of insects (495 individuals representing 19 families and four orders) were recorded, compared with 97 species on LTIs (2018 individuals, 21 families, four orders) (Table 1). Nearly 50% of the species recorded on both STIs and LTIs were Lepidoptera, followed by Hymenoptera (37% and 44% of the insect species recorded on STIs and LTIs, respectively) with Coleoptera and Diptera contributing the remaining taxa. Species represented by ≤ 4 individuals are regarded as rare and those represented by ≤ 5 individuals as common (see supplementary Appendix S1).

Table 1.  Number of species (percentage of species within each taxon) recorded at edges of small and large tree islands in Taylor Slough, Everglades National Park
OrderFamilySmall tree islandsLarge tree islands
Lepidoptera 21 (49%)46 (48%)
Papilionidae 3 (7%) 5 (5%)
Pieridae 5 (12%)10 (10%)
Lycaenidae 2 (5%) 5 (5%)
Riodinidae 1 (2%) 1 (1%)
Nymphalidae 7 (16%)14 (14%)
Danaidae 1 (2%) 3 (3%)
Hesperiidae 2 (5%) 7 (7%)
Arctiidae  1 (1%)
Hymenoptera 16 (37%)43 (44%)
Colletidae 2 (5%) 5 (5%)
Halictidae 1 (2%) 6 (6%)
Megachilidae 4 (9%)12 (12%)
Anthophoridae 3 (7%) 7 (7%)
Apidae 2 (5%) 3 (3%)
Vespidae 2 (5%) 6 (6%)
Sphecidae 2 (5%) 4 (4%)
Coleoptera  4 (9%) 3 (3%)
Cerambycidae 1 (2%) 1 (1%)
Chrysomelidae 1 (2%) 
Scarabidae 2 (5%) 2 (2%)
Diptera  2 (5%) 5 (5%)
Syrphidae 1 (2%) 2 (2%)
Muscidae 1 (2%) 1 (1%)
Stratiomyidae  1 (1%)
Tachinidae  1 (1%) 

Lepidoptera represented 63% and 65% of the total number of individuals captured from STIs and LTIs, respectively (Table 2) and over 60% of these carried A. lanceolata pollinia. Ninety-one per cent of Danaus gilippus Cramer (Queen butterflies) captured and 77% of Papilio spp. (swallowtail butterflies) captured on the edge of tree islands carried pollinia, and two other common butterfly species, Eurema daira Godart (barred yellow) and Phyciodes phaon (Edwards) (Phaon crescent), often carried pollinia (Table 2).

Table 2.  Number of insect visitors recorded during sampling at edges of tree islands and during observations at potted Asclepias lanceolata plants at four distances from tree islands in marl prairie habitat
Order, family and speciesSTILTIDistance from STI (m)Distance from LTI (m)
  • a

    0% of captured insects with pollinia;

  • b


  • c


  • d


  • e


  Papilio polyxenes 5d 6e 2e1e1a 4d 2c
  P. cresphontes 6c2c 3d 2e1e
  P. palamedes 1e 3d 1e1a
  P. glaucus 1e
  Phoebis sennae  1e 2e 1e
  Eurema daira13d32d 5d 1e
  E. nicippe  7d 1a 7d 2c
  Agraulis vanillae 2e 8d 5c2c 5c 4e2c
  Euptoieta claudia 5c10d 4c2c 4e 1a1a
  Phyciodes phaon10c34d 8d5e28d12d
  P. tharos17c17d 7d
  Anartia jatrophae 4c 7c 4b 2c1a
  Junonia coenia 5b17b
  Danaus gilippus 8e25d 5e2c2e20e15e8d1e
  Bombus pennsylvanicus 3c11d 1e1a 2c 1e
  Xylocopa micans 1e 3e 1e 1e
  Polistes metricus 7b22b 3a 8b 6a
  P. fuscatus bellicosus18b33b10b3a21b10b
  Typocerus zebra 2a25a 3c

Individuals in the order Hymenoptera were the second most recorded insect group, accounting for 35% and 26% of total individuals from the edges of STIs and LTIs, respectively (Table 2). All Xylocopa micans Lepeletier (carpenter bees) captured carried pollinia, as did 33% and 64%, respectively, of Bombus pennsylvanicus DeGeer (bumblebees) on STIs and LTIs. Polistes spp. (paper wasps) were the most common hymenopteran taxa captured, but fewer than 20% carried pollinia. One beetle species, Typocerus zebra Olivier, was captured in both STIs and LTIs, but none of the individuals collected carried pollinia.

marl prairies and their pollinator communities

Pollinator abundance and number of species present in the marl prairie declined with increasing distance from tree islands (Fig. 1). Of the 52 species identified along transects radiating into the marl prairie from STIs, 49 were encountered within 25 m of the STIs. All 89 species observed along transects radiating from LTIs were recorded within 25 m, although 55% were found in low abundance (≤ 4 individuals). Lepidoptera and Hymenoptera accounted for 92% and 90% of individuals recorded at 25 m from STIs and LTIs, respectively.


Figure 1. Distribution of insects along transects radiating from small and large bayhead tree islands: (a) species richness and (b) number of individuals. Transect data were pooled for all the islands within the two tree island size classes. Numbers above bars are the number of species (a) or number of individuals (b) at each distance class.

Download figure to PowerPoint

Most of the individuals recorded along transects were noted within 200–300 m of STIs and LTIs (Fig. 1b). Numbers were greatest in the 0–25 m class, with a precipitous decline in pollinator activity with increasing distance away from tree islands.

Strong, long-distance fliers including Xylocopa spp., Bombus pennsylvanicus, Apis mellifera L., Polistes spp., Papilio spp., Agraulis vanillae L., Junonia coenia Hübner and Danaus gilippus were all observed at distances ≥ 700 m from tree islands. However, only Papilio cresphontes Cramer, Euptoieta claudia Cramer and Danaus gilippus were observed visiting Asclepias lanceolata at such distances and, of these, only P. cresphontes and D. gilippus were noted to carry pollinia (Table 2). The mean body size of bees did not change in relation to distance from STIs (F = 2.069, P = 0.07, n = 217; Fig. 2a), but increased with increasing distance from LTIs (F = 4.872, P < 0.001, n = 700; Fig. 2b).


Figure 2. Mean combined thorax and abdomen lengths of bees as a function of capture distance recorded along transects from (a) small tree islands and (b) large tree islands in marl prairie habitat. Means and pooled standard errors (bars represent ± 1 SE) are shown.

Download figure to PowerPoint

A taxonomically diverse pollinator assemblage visited A. lanceolata. A total of 52 insect species were observed visiting flowers of A. lanceolata throughout the study and, of those captured and examined after a floral visit, 17 species were noted to carry pollinia (Table 2). Diurnal Lepidoptera accounted for 81% of the individuals visiting A. lanceolata, while Hymenoptera, Coleoptera and Diptera comprised the remaining taxa around LTIs (only Hymenoptera around STIs). For data pooled across all tree island size and distance classes, Lepidoptera accounted for 93% of the insects with attached pollinia after visiting A. lanceolata, Hymenoptera for 6% and Coleoptera for the remaining 1%.

pollination success of a. lanceolata in the marl prairie

Isolation (distance from tree island) and the size of the neighbouring tree island significantly affected the reproductive success of potted A. lanceolata plants placed in the marl prairie (Fig. 3). The number of fruits produced per plant was greater for plants placed closer to the edges compared with plants placed further away for both STIs (F3,36 = 3.716, P = 0.020; Fig. 3a) and LTIs (F3,44 = 4.889, P = 0.005; Fig. 3b), but twice as many fruits per plant were produced 25 m from LTIs than at the same distance from STIs (F1,20 = 7.68, P = 0.012). There is an effect of distance on the number of seeds produced per plant for both STIs (R2 = 0.198, P = 0.004, n = 40; Fig. 4a) and LTIs (R2 = 0.334, P < 0.001, n = 48; Fig. 4b) with plants at 25 m producing, on average, more than three times as many seeds when adjacent to LTIs rather than STIs (F1,20 = 14.686, P = 0.001). Tree island size also affected the percentage germination of seeds produced by plants at 25 m (F1,25 = 5.615, P = 0.026; Fig. 5) but not at greater distances.


Figure 3. Relationship between distance and reproductive success of Asclepias lanceolata represented as mean fruit production per plant on small (a) and large tree islands (b). Bars with different letters are significantly different at P < 0.05. Bars show means ± 1 SE.

Download figure to PowerPoint


Figure 4. Relationship between distance and reproductive success of Asclepias lanceolata represented as mean seed production per plant on small (a) and large tree islands (b).

Download figure to PowerPoint


Figure 5. Effect of distance from tree islands on mean germination percentage of seeds of Asclepias lanceolata. Error bars represent ± 1 SE. *P < 0.05; NS = P > 0.05.

Download figure to PowerPoint

Species richness of pollinators visiting A. lanceolata depended on both the size of and the distance from the nearest tree island (Fig. 6). More than three times as many individuals were recorded visiting plants at 25 m from LTIs than from STIs. Of the species recorded visiting A. lanceolata≥ 300 m from both STIs and LTIs, 86% were Lepidoptera, all typical open-habitat specialists.


Figure 6. Relationship between degree of isolation from small and large tree islands and number of species. Numbers above columns represent cumulative species richness.

Download figure to PowerPoint


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

tree islandmarl prairie pollinator communities

Our results show that tree island size affects insect pollinator diversity and abundance along the edges of tree islands and in the surrounding marl prairie. LTIs supported a higher abundance and diversity (twice as many species) of insects than STIs, even though the two size classes share many of the same insect taxa. LTIs presumably offer more resistance to environmental factors such as wind, intensity of solar radiation and fluctuations in temperature and thus probably provide more suitable habitat for nesting and roosting pollinators. The flora and fauna associated with tree island communities would be expected, according to island biogeography theory, to exhibit increasing species diversity with increasing island size and decreased isolation (MacArthur & Wilson 1967; Simberloff 1976). In this study, not only are the tree island communities and their associated plants and insect pollinators surrounded by an expansive marl prairie, which provides resources (e.g. nectar and pollen, nesting materials, larval host plants) to the tree island insects, but insect-pollinated plants in the marl prairie also depend on this important source of potential pollinators. Thus, both the size of tree islands and their spatial arrangement in the landscape can affect the interactions within the matrix, as seen in the effect of isolation from different size tree islands on pollination success of plants in the matrix.

Distance from tree islands, as well as their size, had a pronounced effect on the distribution of insect species and abundance in the marl prairie habitat. Pollinator abundance and number of species declined with increasing distance away from tree islands, which we ascribe, at least in part, to the dispersal and flight capabilities of pollinators.

Characteristically strong-flying pollinators were noted at distances up to 1 km, although, of the 10 such species, only Papilio cresphontes and Apis mellifera were observed ≥ 5 times during this study. Likewise, of those species captured and examined for pollinia attachment at 1 km, seven were recorded to carry pollinia, with P. cresphontes and Euptoieta claudia noted to carry multiple pollinia clusters. These findings suggest that plants in isolated areas that rely on insects for pollination services may be at a disadvantage if they depend on dispersal-limited, weak-flying flower specialists.

Both STIs and LTIs provided nesting substrate, larval host plants and shelter for a taxonomically diverse guild of potential pollinators. Many butterfly species were commonly seen foraging for nectar, mating, ovipositing and basking at tree island edges throughout the study (D. Artz, personal observation). Certain species of swallowtail butterflies have particular host and oviposition preferences for plants only found in tree islands: Papilio palamedes oviposit only on Persea borbonia, whereas larvae of P. glaucus feed on Magnolia virginiana (Scriber et al., 2000). Although Papilio palamedes and P. glaucus were not as common as other swallowtail butterflies in the landscape, both species were occasionally observed ovipositing on Persea borbonia (D. Artz, personal observation).

Bayhead tree islands also supported a diversity of hymenopteran taxa, with an assemblage comprised predominantly of Bombus pennsylvanicus, Xylocopa spp., Apis mellifera, Megachile spp. and Polistes spp. Although no honeybee colonies were discovered in any of the tree islands studied, a long-term study of honeybee populations within ENP shows that they nest in various habitats, including tree islands (our unpublished data).

pollination and visitor abundance

A wide taxonomic diversity of potential insect pollinators visited Asclepias lanceolata. Similar diversity of insects has been recorded in Asclepias incarnata, A. syriaca, A. tuberosa and A. verticillata (Kephart 1983; Wyatt & Broyles 1994; Ivey et al. 2003; Kephart & Theiss 2004). There is evidence that Bombus and Apis were the most effective pollinators of A. tuberosa in single-visit pollen transfer, but lepidopteran taxa were the most abundant visitors and typically carried more pollinia (Fishbein & Venable 1996). Asclepias incarnata was predominately pollinated by hymenopteran taxa, which carried more pollinia than any other insect order on average, with carpenter bees (Xylocopa virginica) carrying three times as many A. incarnata pollinia as any other insect species (Ivey et al. 2003).

The visitors to A. lanceolata probably differ in efficiency as pollinators because they differ in ability to remove and deposit pollinia. Pollination efficiency can be affected by the morphology and overall ability of the foraging insect to remove pollinia from flowers (Inouye et al. 1994). We observed two Eurema daira individuals dead on flowers of A. lanceolata, presumably unable to dislodge their tarsal appendages from the pollinia apparatus. In addition, incidences of nectar robbing at flowers of A. lanceolata by Xylocopa spp. and Polistes spp. reduce the likelihood of legitimate pollination by these hymenopteran taxa. Papilio spp., Danaus gilippus and Phyciodes phaon had a high proportion of captured individuals that carried pollinia. Willson & Bertin (1979) found that lepidopterans accounted for 25–40% of insect vectors of A. syriaca pollinia at four study sites. Based on pollinator visitation and capture data and those with a high percentage of attached pollinia, lepidopteran taxa are a potentially effective group of pollinators of A. lanceolata.

Although other studies have shown that nocturnal moths are effective pollinators of milkweeds (Bertin & Willson 1980; Jennersten & Morse 1991), investigations of pollinia removal and insertion in A. lanceolata flowers by nocturnal moths suggest that they are unlikely to be pollinators of this species (D. Artz, unpublished data). However, because we did not routinely sample insect visitors at night, there is potential that some nocturnal species such as geometrid and noctuid moths may have visited A. lanceolata flowers.

Experimental milkweed plants were visited at distances up to 1 km from tree islands by several insect taxa, many of them open-habitat specialists, but other taxa were not found so far from tree islands, suggesting that differences in the movement of individuals in diverse habitats affect the pollination success of plants in isolated areas. Similar results have been reported for other studies on foraging behaviour in bumblebees and honeybees (Walther-Hellwig & Frankl 2000; Kreyer et al. 2004). Small-bodied solitary bees, however, were recorded at distances only up to 400 m away from tree islands with most of the individuals observed at distances of less than 200 m, consistent with measured foraging ranges of several species of European solitary bees at between 150 m and 600 m (Gathmann & Tscharntke 2002). Our data provide further evidence that small-bodied solitary bees have limited foraging ranges and, in addition, that foraging distance is positively correlated with body size.

The observed pattern of pollinator density near tree islands and the decline in abundance with increasing distance away from tree islands can also be attributed to the foraging ecology and behaviour of these insects. Most bees and some butterflies are central-place foragers such that, after foraging bouts, the insects return to their nest or refuge (Cresswell et al. 2000). Moreover, some central-place foragers exhibit territorial behaviour to protect their nests or to defend their mating sites (Orians & Pearson 1979). In this study, male carpenter bees (Xylocopa micans) and a few butterfly species (e.g. Limenitis archippus Cramer) were occasionally observed exhibiting territorial behaviour in and around tree islands. Such foragers search for floral resources near refugia or the nest site to enhance the energetic rewards and reduce travel costs and to reduce predation risk in the open habitat. The observed pattern of pollinator density near tree islands corroborates this hypothesis. For both STIs and LTIs, the number of recorded individuals was greatest at 25 m away from tree islands and declined precipitously with increasing distance, suggesting that overall pollinator activity was greatest near tree islands.

Distance from STIs and LTIs significantly influenced the pollination success of Asclepias lanceolata; seed set varied inversely with distance from tree islands. One caveat about interpretation of the results of seed set from STIs is that the regression line is strongly affected by one data point at 1000 m. However, the overall trend in seed set does decrease with increasing distance. Similar results are reported in other studies examining the effect of isolation on plant reproductive success in other natural habitats and agro-ecosystems (Steffan-Dewenter & Tscharntke 1999; Wolf & Harrison 2001). The effects of distance from grassland and agricultural patches on plant reproductive success and pollinator abundance in experimentally created ‘habitat islands’ resulted in a decrease in diversity of flower-visiting insects with increasing isolation. Seed production per plant, however, significantly decreased with increasing isolation (Steffan-Dewenter & Tscharntke 1999).

Our study is the first to show effects of tree island pollinator diversity and abundance, tree island size and distance from tree islands on plant–animal interactions in the matrix, rather than simply on biotic interactions within tree islands. We ascribe the differential responses of insect pollinators, such as foraging behaviour and flight distance, to the spatial structure and complexity of the landscape, particularly tree island size and tree island density. These findings have fundamental implications for understanding reproduction in plants that require pollinator services, particularly in small, isolated populations. Tree islands in the Everglades landscape support a diverse assemblage of entomofauna by providing refuge habitats and nesting sites, but also serve as a source of potential pollinators for the surrounding marl prairie plant communities. Our study illustrates the need for further understanding of species interactions in other mosaic landscapes.


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

We thank Stewart Schultz for advice on statistical procedures and Colin Hughes and Suzanne Koptur for advice on experimental design and analysis. We are grateful to Carlos Garcia-Robledo, Carol Horvitz, Colin Hughes, Suzanne Koptur and Mindy Nelson for helpful comments and suggestions on the manuscript. We thank the personnel at Everglades National Park for permitting the research in ENP. Troy Mullins assisted selection of tree islands using ENP GIS maps. The work was funded, in part, by a J. Gerry Curtis Plant Science Scholarship from the Department of Biology at the University of Miami to D.R.A.


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information
  • Ågren, J. (1996) Population size, pollinator limitation, and seed set in the self-incompatible herb Lythrum salicaria. Ecology, 77, 17791790.
  • Aizen, M.A. & Feinsinger, P. (1994) Forest fragmentation, pollination, and plant reproduction in a Chaco dry forest, Argentina. Ecology, 75, 330351.
  • Bawa, K.S. (1990) Plant–pollinator interactions in tropical rain forests. Annual Review of Ecology and Systematics, 21, 399422.
  • Beekman, M. & Ratnieks, F.L.W. (2000) Long range foraging in the honeybee Apis mellifera. Functional Ecology, 14, 490496.
  • Benitez-Malvido, J., Garcia-Guzman, G. & Kossmann-Ferraz, I.D. (1999) Leaf-fungal incidence and herbivory on tree seedlings in tropical rainforest fragments: an experimental study. Biological Conservation, 91, 143150.
  • Bertin, R.I. & Willson, M.F. (1980) Effectiveness of diurnal and nocturnal pollination of two milkweeds. Canadian Journal of Botany, 58, 17441746.
  • Borror, D.J., Triplehorn, C.A. & Johnson, N.F. (1997) An Introduction to the Study of Insects, 6th edn. Saunders College Publishing, Philadelphia.
  • Bronstein, J.L. (1995) The plant–pollinator landscape. Mosaic Landscapes and Ecological Processes (eds L. Hansson, L. Fahrig & G. Merriam), pp. 257288. Chapman & Hall, London.
  • Cresswell, J.E., Osborne, J.L. & Goulson, D. (2000) An economic model of the limits to foraging range in central place foragers with numerical solutions for bumblebees. Ecological Entomology, 25, 249255.
  • Davis, S.M. & Ogden, J.C. (1994) Everglades: the Ecosystem and its Restoration. St. Lucie Press, Delray Beach, Florida.
  • Debinski, D.M., Ray, C. & Saveraid, E.H. (2001) Species diversity and the scale of the landscape mosaic: do scales of movement and patch size affect diversity? Biological Conservation, 98, 179190.
  • Fishbein, M. & Venable, D.L. (1996) Diversity and temporal change in the effective pollinators of Asclepias tuberosa. Ecology, 77, 10611073.
  • Forman, R.T.T. (1995) Land Mosaics: the Ecology of Landscapes and Regions. Cambridge University Press, Cambridge.
  • Gathmann, A. & Tscharntke, T. (2002) Foraging ranges of solitary bees. Journal of Animal Ecology, 71, 757764.
  • Gunderson, L.H. (1994) Vegetation of the Everglades: determinants of community composition. Everglades: the Ecosystem and its Restoration (eds S.M. Davis & J.C. Ogden), pp. 323340. St. Lucie Press, Delray Beach, Florida.
  • Haddad, N.M. (1999) Corridor use predicted from behaviors at habitat boundaries. American Naturalist, 153, 215227.
  • Haynes, K.J. & Cronin, J.T. (2004) Patch quality is too often ignored in landscape-matrix studies. Landscape Ecology, 19, 119124.
  • Heinrich, B. (1979) Resource heterogeneity and patterns of foraging in bumblebees. Oecologia, 40, 234245.
  • Hirsch, M., Pfaff, S. & Wolters, V. (2003) The influence of matrix type on flower visitors of Centaurea jacea L. Agriculture, Ecosystems and Environment, 98, 331337.
  • Inouye, D.W., Gill, D.E., Dudash, M.R. & Fenster, C.B. (1994) A model and lexicon for pollen fate. American Journal of Botany, 81, 15171530.
  • Ivey, C.T., Martinez, P. & Wyatt, R. (2003) Variation in pollinator effectiveness in swamp milkweed, Asclepias incarnata (Apocynaceae). American Journal of Botany, 90, 214225.
  • Janzen, D.H. (1971) Euglossine bees as long-distance pollinators of tropical plants. Science, 171, 203205.
  • Jennersten, O. & Morse, D.H. (1991) The quality of pollination by diurnal and nocturnal insects visiting common milkweed, Asclepias syriaca. American Midland Naturalist, 125, 1828.
  • Kephart, S.R. (1983) The partitioning of pollinators among three species of Asclepias. Ecology, 64, 120133.
  • Kephart, S.R. & Theiss, K. (2004) Pollinator-mediated isolation in sympatric milkweeds (Asclepias): do floral morphology and insect behavior influence species boundaries? New Phytologist, 161, 265277.
  • Krebs, C.J. (1989) Ecological Methodology. Harper & Row, New York.
  • Kreyer, D., Oed, A., Walther-Hellwig, K. & Frankl, R. (2004) Are forest potential landscape barriers for foraging bumblebees? Landscape scale experiments with Bombus terrestris agg. and Bombus pascuorum (Hymenoptera, Apidae). Biological Conservation, 116, 111118.
  • Kushlan, J.A. (1990) Freshwater marshes. Ecosystems of Florida (eds R.L. Myers & J.J. Ewel), pp. 324363. University of Central Florida Press, Orlando, Florida.
  • Lodge, T.E. (1994) The Everglades Handbook: Understanding the Ecosystem. St. Lucie Press, Delray Beach, Florida.
  • MacArthur, R.H. & Wilson, E.O. (1967) The Theory of Island Biogeography. Princeton University Press, Princeton, New Jersey.
  • Magurran, A.E. (1988) Ecological Diversity and its Measurement. Princeton University Press, Princeton, New Jersey.
  • Masaki, T. (2004) Effect of the shape of forest fragments on tree population dynamics. Plant Ecology, 172, 275286.
  • McEuen, A.B. & Curran, L.M. (2004) Seed dispersal and recruitment limitation across spatial scales in temperate forest fragments. Ecology, 85, 507518.
  • Michener, C.D. (2000) The Bees of the World. Johns Hopkins University Press, Baltimore, Maryland.
  • Michener, C.D., McGinley, R.J. & Danforth, B.N. (1994) The Bee Genera of North and Central America. Smithsonian Institution Press, Washington, DC.
  • Minno, M.C. & Emmel, T.C. (1993) Butterflies of the Florida Keys. Scientific Publishers, Gainesville, Florida.
  • Mitchell, T.B. (1960) Bees of the Eastern United States, Vol. 1. Technical Bulletin 141. North Carolina Agricultural Experiment Station, North Carolina.
  • Mitchell, T.B. (1962) Bees of the Eastern United States, Vol. 2. Technical Bulletin 152. North Carolina Agricultural Experiment Station, North Carolina.
  • Moertberg, U.M. (2001) Resident bird species in urban forest remnants: landscape and habitat perspectives. Landscape Ecology, 16, 193203.
  • Olmsted, I.C., Loope, L.L. & Rintz, R.E. (1980) A Survey and Baseline Analysis of Aspects of the Vegetation of Taylor Slough, Everglades National Park. Report T-586, South Florida Research Center. National Park Service, US Department of the Interior, Homestead, Florida.
  • Orians, G.H. & Pearson, N.E. (1979) On the theory of central-place foraging. Analysis of Ecological Systems (eds D.J. Horn, R.D. Mitchell & G.R. Stairs), pp. 154177. Ohio State University Press, Columbus, Ohio.
  • Osborne, J.L., Clark, S.J., Morris, R., Williams, I., Riley, R., Smith, A., Reynolds, D. & Edwards, A. (1999) A landscape-scale study of bumble bee foraging range and constancy, using harmonic radar. Journal of Applied Ecology, 36, 519533.
  • Perfecto, I. & Vandermeer, J. (2002) Quality of agroecological matrix in a tropical montane landscape: ants in coffee plantations in southern Mexico. Conservation Biology, 16, 174182.
  • Pollard, E. (1977) A method for assessing changes in abundance of butterflies. Biological Conservation, 12, 115134.
  • Pollard, E. & Yates, T.J. (1993) Monitoring Butterflies for Ecology and Conservation. The British Butterfly Monitoring Scheme. Institute of Terrestrial Ecology and Joint Nature Conservation Committee. Chapman & Hall, London.
  • Ries, L. & Debinski, D.M. (2001) Butterfly responses to habitat edges in the highly fragmented prairies of Central Iowa. Journal of Animal Ecology, 70, 840852.
  • Santos, T. & Telleria, J.L. (1994) Influence of forest fragmentation on seed consumption and dispersal of spanish juniper, Juniperus thurifera. Biological Conservation, 70, 129134.
  • Schneider, C. & Fry, G.L.A. (2001) The influence of landscape grain size on butterfly diversity in grasslands. Journal of Insect Conservation, 5, 163171.
  • Scriber, J.M., Margraf, N. & Wells, T. (2000) Suitability of four families of Florida ‘Bay’ species for Papilio palamedes and P. glaucus (Papilionidae). Journal of the Lepidopterists’ Society, 54, 131136.
  • Simberloff, D. (1976) Experimental zoogeography of islands: effects of island size. Ecology, 57, 629648.
  • Sklar, F.H. & van der Valk, A.G. (2002) Tree Islands of the Everglades. Kluwer Academic Publishers, Dordrecht, The Netherlands.
  • Steffan-Dewenter, I. & Kuhn, A. (2003) Honeybee foraging in differentially structured landscapes. Proceedings of the Royal Society of London, 270, 569575.
  • Steffan-Dewenter, I. & Tscharntke, T. (1999) Effects of habitat isolation on pollinator communities and seed set. Oecologia, 121, 432440.
  • Summerville, K.S. & Crist, T.O. (2003) Determinants of lepidopteran community composition and species diversity in eastern deciduous forests: roles of season, eco-region and patch size. Oikos, 100, 134148.
  • Sutcliffe, O.L. & Thomas, C.D. (1996) Open corridors appear to facilitate dispersal by ringlet butterflies (Aphantopus hyperantus) between woodland clearings. Conservation Biology, 10, 13591365.
  • Tewksbury, J.J., Levey, D.J., Haddad, N.M., Sargent, S., Orrock, J.L., Weldon, A., Danielson, B.J., Brinkerhoff, J., Damschen, E.I. & Townsend, P. (2002) Corridors affect plants, animals, and their interactions in fragmented landscapes. Proceedings of the National Academy of Sciences, 99, 1292312926.
  • Tscharntke, T., Steffan-Dewenter, I., Kruess, A. & Thies, C. (2002) Characteristics of insect populations on habitat fragments: a mini review. Ecological Research, 17, 229239.
  • Visscher, P.K. & Seeley, T.D. (1982) Foraging strategy of honeybee colonies in a temperate deciduous forest. Ecology, 63, 17901801.
  • Waddington, K.D., Visscher, P.K., Herbert, T.J. & Richter, M.R. (1994) Comparisons of forager distributions from matched honey bee colonies in suburban environments. Behavioral Ecology and Sociobiology, 35, 423429.
  • Walther-Hellwig, K. & Frankl, R. (2000) Foraging habitats and distances and foraging distances of bumblebees, Bombus spp. (Hym., Apidae), in an agricultural landscape. Journal of Applied Entomology, 124, 299306.
  • Wiens, J.A. (1997) Metapopulation dynamics and landscape ecology. Metapopulation Biology: Ecology, Genetics, and Evolution (eds I.A. Hanski & M.E. Gilpin), pp. 4367. Academic Press, San Diego.
  • Willson, M.F. & Bertin, R.I. (1979) Flower-visitors, nectar production, and inflorescence size of Asclepias syriaca. Canadian Journal of Botany, 57, 13801388.
  • Wolf, A.T. & Harrison, S.P. (2001) Effects of habitat size and patch isolation on reproductive success of the serpentine morning glory. Conservation Biology, 15, 111121.
  • Wyatt, R. & Broyles, S.B. (1994) Ecology and evolution of reproduction in milkweeds. Annual Review of Ecology and Systematics, 25, 423441.

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Appendix S1. List of insect species recorded at tree island edges in Taylor Slough, Everglades National Park.

JEC_1109_sm_AppS1.doc115KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.