Figures
Abstract
Despite increasing knowledge about the effects of habitat loss on pollinators in natural landscapes, information is very limited regarding the underlying mechanisms of forest fragmentation affecting plant-pollinator interactions in such landscapes. Here, we used a network approach to describe the effects of forest fragmentation on the patterns of interactions involving the understory dominant palm Astrocaryum mexicanum (Arecaceae) and its floral visitors (including both effective and non-effective pollinators) at the individual level in a Mexican tropical rainforest landscape. Specifically, we asked: (i) Does fragment size affect the structure of individual-based plant-pollinator networks? (ii) Does the core of highly interacting visitor species change along the fragmentation size gradient? (iii) Does forest fragment size influence the abundance of effective pollinators of A. mexicanum? We found that fragment size did not affect the topological structure of the individual-based palm-pollinator network. Furthermore, while the composition of peripheral non-effective pollinators changed depending on fragment size, effective core generalist species of pollinators remained stable. We also observed that both abundance and variance of effective pollinators of male and female flowers of A. mexicanum increased with forest fragment size. These findings indicate that the presence of effective pollinators in the core of all forest fragments could keep the network structure stable along the gradient of forest fragmentation. In addition, pollination of A. mexicanum could be more effective in larger fragments, since the greater abundance of pollinators in these fragments may increase the amount of pollen and diversity of pollen donors between flowers of individual plants. Given the prevalence of fragmentation in tropical ecosystems, our results indicate that the current patterns of land use will have consequences on the underlying mechanisms of pollination in remnant forests.
Citation: Dáttilo W, Aguirre A, Quesada M, Dirzo R (2015) Tropical Forest Fragmentation Affects Floral Visitors but Not the Structure of Individual-Based Palm-Pollinator Networks. PLoS ONE 10(3): e0121275. https://doi.org/10.1371/journal.pone.0121275
Academic Editor: Stephen C. Pratt, Arizona State University, UNITED STATES
Received: September 15, 2014; Accepted: January 29, 2015; Published: March 31, 2015
Copyright: © 2015 Dáttilo et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited
Data Availability: All relevant data are within the paper.
Funding: The authors received no specific funding for this work.
Competing interests: The authors have declared that no conflict of interest exists.
Introduction
Deforestation is the major cause of fragmentation and spatial isolation of populations in the tropics [1], with potential to increase species or population extinction rates [1,2]. Despite the importance of ecological interactions in the structure and stability of biological communities over time and space [3], most studies have focused only on populations and species loss, but studies on the consequences of fragmentation on ecological processes are gaining prominence in the literature [4, 5, 6, 7]. Indeed, recent research indicates that species interactions are increasingly at risk of local and global extinction as a consequence of changes in land use [7, 8, 9].
Fragmentation is one of the major threats to plant-pollinator interactions in the tropics [7, 10, 11]. This occurs mainly because the loss of habitat decreases pollinator activity, pollen deposition, and outcrossing levels, which directly influences the reproductive success of plants remaining in isolated fragments [7, 12, 13]. Recent studies have also suggested that there is a global pollination crisis due to the decline in both abundance and richness of pollinators [14, 15]. Under a unified framework of species interactions, studies have focused on the structural properties of plant-pollinator networks at the community level in different systems and habitats throughout the Earth [16, 17, 18, 19], including the effects of habitat loss on these networks [20, 21, 22]. The topological structure of these ecological networks can be described through patterns of species interactions such as nestedness, modularity, specialization and diversity of interactions (reviewed by [23]). Nestedness describes the degree to which specialists interact with a subset of the species interacting with generalists. Modularity describes the presence of semi-independent compartments of highly interacting species. Specialization refers to the overall level of dependence of all interacting species in a network. Diversity of interactions describes the magnitude of the interactions between plants and animals. Despite the increase in knowledge on plant–pollinator networks at the community level, few studies have evaluated how these mutualistic relationships vary within biological populations, but see [24].
We know that within a given population of a plant species, there is variation in the frequency of flower visitation within and among individual plants [25, 26, 27]. One of the main possible factors explaining such intrapopulation variation is based on differences in the quality of reward, and therefore it would be expected that individuals with better rewards would be most visited, ensuring cross-pollination. However, plants with longer flowering time could also be the most visited. Consequently, such inter-individual variation can directly affect the fitness of plants, since it is expected that the most visited individuals will have greater reproductive success (seed and fruit set) [26, 27]. Variation in pollinator visitation frequency to flowers can generate complex networks among individuals of the same plant species. Similar studies have been conducted involving individual-based networks in other interaction systems using an intrapopulation approach, such as sea otters-prey, didelphid marsupials-arthropods, snails and fruits, monkeys-fruiting plants, frogs, ant-plants, and lizards-diets [28, 29, 30, 31, 32, 33]. However, to our knowledge only two studies have evaluated the networks of plant-pollinator interactions among individuals within populations [26, 27].
In tropical rainforests, palms are an important component of the understory [34, 35, 36]. It is estimated that there are between 2400 [37] and 2700 palm species [38] in tropical and subtropical regions around the world, yet little is known about pollination interactions and reproduction of these plants in fragmented landscapes [39]. Since each palm tree can be visited by a rich contingent of different insect groups [40, 41], we examined if these plant-insect interactions could be a good model to study individual-based pollination networks. To our knowledge, this is the first study that directly evaluates intrapopulation variations in pollinator activity of palms in fragmented landscapes. Here, we used a network approach to describe the effects of forest fragmentation on the patterns of interactions involving the understory dominant palm Astrocaryum mexicanum (Arecaceae) and its floral visitors. For this, we collected effective (Coleoptera: Nitidulidae) and non-effective pollinators on male- and female-phase inflorescences in populations of A. mexicanum (Arecaceae) in different tropical rainforest fragments of Los Tuxtlas, in southeast Mexico. We hypothesized that the size of the forest fragment could affect the abundance of effective and non-effective pollinators of A. mexicanum, and consequently, influence the pattern of interactions within the networks. We then addressed the following questions: (i) Does fragment size affect the structure of individual-based plant-pollinator networks? (ii) Does the core of highly interacting visitor species change along the fragmentation size gradient? (iii) Does forest fragment size influence the abundance of effective pollinators of A. mexicanum?
Materials and Methods
Ethics statement
All research was conducted in accordance with national and international guidelines, and conforms to the legal requirements of the Mexican government. Los Tuxtlas Biological Station-UNAM granted permission to conduct the study in all selected fragments.
Study site and fragments
This study was conducted at the Los Tuxtlas Biological Station, a field station of the National University of Mexico (UNAM) (18°34′–18°36′ N, 95°04′–95°09′ W) located in the State of Veracruz, Mexico. The area has a complex topography with elevation ranging from sea level to 1650 m above sea level within a short distance [42]. The predominant land use in the area is conversion of forest to cattle grasslands, and rates of deforestation in the decades of the 60’s-late 80’s were ca. 4% per year [43]. Current deforestation rates have decreased because forest remnants are restricted to the most inaccessible areas of the region [44]. Floristic diversity in the area includes a total of 950 known vascular plants [45], which largely consist of plants of Neotropical origin, but higher elevation sites include combinations of Neotropical and Nearctic taxa [42]. Palms are a very important component of the tropical rainforest of Los Tuxtlas [45] with 13 species present [46], but only a few of them, such as Astrocaryum mexicanum, Chamaedorea pinnatifrons, Ch. alternas, Ch. Concolor, Ch. tepejilote and Bactris mexicana attain high population densities [45, 47].
In this study, we sampled six forest fragments encompassing a size-range of 2, 4, 19.4, 34.6, 114.6 and 700 ha, the latter representing the reference site, as it extends along the entire elevational range, up to the maximum elevation in the area. Isolated fragments were selected based on the following characteristics: i) they bore the same vegetation type, tropical rainforest, and therefore were located within a restricted elevation (0–150 m a.s.l.); ii) similar age (<50 years of separation); and iii) all plants were located within a plot of 600 m2 established at the center of each forest fragment and continuous forest to avoid edge effects [39, 48].
Species studied and data collection
Astrocaryum mexicanum is the only species of the genus in Mexico [33], reaching densities of up to 1,000 individuals (> 1 m tall) per hectare at Los Tuxtlas [44, 48]. Plants are monoecious (dichogamous), with up to five inflorescences per plant in both sexual phases of the inflorescences [49] (Fig. 1A and 1B). The effective pollinators are small nitidulid beetles (Mystrops sp., M. mexicanus, Eumystrops centralis and Coleopterus aberrans) (Fig. 1C), however, many other animal species can visit the flowers of A. mexicanum [47, 49]. Previous studies have shown that A. mexicanum supports a great diversity of floral visitors (10 arthropod orders, 60 species), and that habitat fragmentation negatively affected the abundance of the main pollinators, but did not have an effect on the plant´s reproductive success [39]. The abundance of potential pollinators (Coleoptera: Nitidulidae) is striking, ranging from 1270 individuals/plant in small fragments (2ha) to 9,893 individuals/plant in the largest fragment (700 ha) [39]. Within each fragment we established a permanent plot of 600 m2 to collect the inflorescences, and we tagged all reproductive plants of A. mexicanum. From these we selected a group of reproductive plants of similar height (2 m) and age (~40–50 years) to sample flower visitors within each fragment. Visitors to the inflorescences were collected during the peak flowering season (March-May) of 1999. Within each fragment we collected the inflorescences of 10 and 5 randomly selected palms in the female and male sexual phases, respectively, when the activity of flower visitors is more intense [39]. Inflorescences were collected between 0700 and 0800 AM, the time at which we observed the most intense insect activity, according with observations by Búrquez et al. [50].
Female flowers (C) located inside an inflorescence surrounded by rachillae, and one of the main pollinators, the beetle Coleopterus aberrans (Be).
Data analysis
Initially, we used a paired t-test (paired per fragment) to evaluate the differences in the number of floral visitor species and the abundance of effective pollinators between male and female inflorescences. To evaluate the effects of forest fragmentation on the patterns of interactions involving Astrocaryum mexicanum and their floral visitors, we used a network approach and defined each of our six fragments as an adjacency matrix A (or interaction network), where aij = number of interactions from an individual plant j by the floral visitor species i, and zero for absence of interactions [51]. Therefore, we built six different interaction networks according to size of the forest fragment.
To evaluate the specialization of our visitation networks based on quantitative data (i.e., interaction frequencies), we used the specialization index (H2’) [52]. In this index, extreme generalization is H2’ = 0 and extreme specialization is H2’ = 1. By incorporating frequency data, this specialization index is extremely robust to changes in sampling intensity and the number of interacting species compared to most other specialization indices (see more details in: [52,53]. We also calculated the diversity of interactions (ID) through a widely used index for analyses of interaction networks [54, 55, 56, 57].
We also performed a second approach that involves the search for non-random patterns of plant-floral visitor interactions. We estimated nestedness for each network using the NODF-metric [58] in ANINHADO [59]. Values of this metric range from 0 (non-nested) to 100 (perfectly nested). In addition, we tested whether within each network there were groups of floral visitors strongly associated with a particular set of individual plants. For this we used the modularity index (M) based on Simulated Annealing (SA) (range 0–1) [60, 61] using the software MODULAR [62]. This index ranges from 0 = no subgroups, to 1 = totally separated subgroups [16]. Although this index M is used for unipartite networks, our null models control any potential effects of bipartite structure on modularity (interactions only occur between floral visitors and plants) [29, 63]. We tested the significance of NODF and M for each network through 1000 simulated networks generated by the Null Model II (CE) to assess whether the observed values in the empirical networks were higher than would be expected for null distributions of these values [64]. In this null model, the probability of an interaction occurring is proportional to the number of interactions of both floral visitors and plant individuals [51]. We used these network descriptors and null model because they provide a way to characterize the organization of these networks in a way that allows direct comparison with previous work on individual-based networks. We defined floral visitor species as core (i.e., those with the highest number of interactions) or peripheral (i.e., those with the fewest number of interactions) components of the network according to [65].
To explore the effects of forest fragment size on plant-floral visitor networks, we used simple linear regression with the log10-transformed fragment size as a predictor variable for the following network descriptors: richness of floral visitors, network specialization, interaction diversity, nestedness and modularity. We also used simple linear regression to evaluate the effects of forest fragment size on the abundance of effective pollinators on male and female inflorescences.
We tested for differences in peripheral and generalist core species among fragments through a permutation test (10,000 permutations) using Analysis of Similarities (ANOSIM) based on Bray-Curtis’s dissimilarity index (quantitative data). All data analyses were performed using R–software version 3.1.2 [66]. All data are included within the manuscript.
Results
Across all fragments, we collected a total of 228,772 arthropods (10 orders, 60 species or morphospecies) visiting inflorescences of A. mexicanum individuals (considering both male and female phases) (Fig. 2). Coleoptera was the predominant group (> 50% of the species) including the four main pollinators: Eumystrops centralis, Mystrops mexicanus, Coleopterus aberrans, and Mystrops sp. (Coleoptera), followed by Hymenoptera (20%), while the remaining (30%) was distributed among the other eight orders. The average number of floral visitors per individual of A. mexicanum was 34.83 species (range from: 30 to 42 species). Of the 60 floral visitor species collected, just one to eight species were present in the central core of highly generalist species: E. centralis (Coleoptera), M. mexicanus (Coleoptera), C. aberrans (Coleoptera), Mystrops sp. (Coleoptera), Unidentified sp1 (Diptera), Unidentified sp1 (Coleoptera), Unidentified sp2 (Coleoptera). Male inflorescences had higher numbers of floral visitor species (Mean ± SD: 18.52 ± 1.93 species) when compared with female inflorescences (12.49 ± 0.79 species) (t = −3.8; df = 5; p = 0.01). However, the abundance of pollinators per female inflorescence was higher (530.63 ± 813.04) compared to male inflorescences (425.31 ± 530.63) (t = −2.755; df = 5; p = 0.04).
The right nodes represent different individuals of A. mexicanum considering both (M) male- and (F) female-phase inflorescence. The left nodes correspond to species of floral visitors that interact with plant individuals. Lines indicate interactions between the two trophic levels. Networks were ordered by both number of links and interaction frequencies. Rectangle height is proportional to the number of interactions recorded per species. Different line lengths indicate the frequency of interactions.
When we evaluated the effect of forest fragmentation on the interaction networks, we found that the number of visitor species within each network did not change regardless of the fragment size (Mean ± SD: 34.83 ± 4.21 species. r2 = 0.042, f = 0.176, df = 1, p = 0.697) (Fig. 3A). In general, we found a low level of network specialization (H2’ = 0.16 ± 0.11) and no trend in variation in the specialization values across fragment sizes (r2 = 0.114, f = 0.023, df = 1, p = 0.886) (Fig. 3B.) The mean interaction diversity index in our networks was 3.71 (± 0.17) and also remained stable over the fragments studied (r2 = 0.069, f = 0.297, df = 1, p = 0.615) (Fig. 3C). We evaluated non-random patterns of species interaction within each network and found that the six networks exhibited a significantly nested network topology (all p-values < 0.01), with NODF-values ranging from 56.68 to 64.54. In this case, we observed that plant individuals with higher number of interactions were visited by both effective pollinators and non-pollinators while individuals with few interactions were only visited by pollinators, generating the cohesive subgroups expected in nested networks. On the other hand, no network was significantly modular when compared with the neutral patterns of interactions (null models) (Modularity values ranged from 0.151 to 0.208.) Both nestedness and modularity values did not vary according to the size of the fragments (Nestedness: r2 = 0.217, f = 1.109, df = 1, p = 0.352. Modularity: r2 = 0.027, f = 0.110, df = 1, p = 0.757) (Fig. 3D and 3E). In addition, we observed that abundance and the standard deviation of effective pollinators per male and female inflorescence of A. mexicanum significantly increased with forest fragment size (male: r2 = 0.76, p = 0.02; female: r2 = 0.72, p = 0.03) (Fig. 4).
Regression correlation coefficient (r) and significance (p) computed using simple linear regressions are also shown.
Regression correlation coefficient (r) and significance (p) computed using simple linear regressions are also shown.
We observed that the richness of flower visitor species changed within each network (from six to eight species). However, richness of flower visitor species was not related to the size of fragments (r2 = 0.154, f = 0.729, df = 6, p = 0.441) (Fig. 3F). Finally, although the species composition of floral visitors found in the network periphery changed between fragments (ANOSIM: r = 0.208, p<0.001), the central core of highly generalist visitor species remained stable (ANOSIM: r = 0.021, p = 0.123).
Discussion
In this study, standardizing the sampling effort based on male- and female-phase inflorescences, we empirically showed for the first time that individual-based palm-pollinator networks are highly nested and non-modular and the sizes of the fragments studied did not affect the topological structure of these networks. Moreover, although the network periphery (i.e, those with fewer interactions) of non-effective pollinators changed across fragments, core generalist species of effective pollinators remained stable over the spatial scale studied. We also observed that both abundance and variance of effective pollinators per male and female inflorescence of A. mexicanum increased with forest fragment size. Theses findings indicate that the presence of effective pollinators in the core of all forest fragments could keep the structure of the networks stable along the gradient of forest fragmentation. However, pollination of A. mexicanum could be more effective in larger fragments, since the greater abundance of pollinators in these fragments could increase the amount of pollen and diversity of pollen donors transferred between flowers.
In individuals of A. mexicanum, anthesis of male and female inflorescences occurs in local asynchrony [50]. Interestingly, female inflorescences open their flowers at 05:00 AM and increase the temperature inside the inflorescence up to 13.5°C (reaching a maximum of 28°C) [50]. Based on these two mechanisms (dichogamy and thermogenesis), we should expect female inflorescences to have a greater richness of floral visitors. However, we observed that male inflorescences have higher richness of visitors per individual plant in all fragments studied. Thermogenesis is an interesting mechanism found in some angiosperm families including Arecaceae [67, 68]. The ecological implications of thermogenesis on pollination in palms are diverse and unknown in most cases, but it has been associated with 1) pollen growth tube, 2) stimulation to pollinators to leave the inflorescences, 3) diffusion of volatiles, and 4) an increase of temperature as a growth-promoting factor to eggs and larvae [68]. According to Búrquez et al. [50], most visitors of A. mexicanum concentrate their activities in two types of rewards: feeding on the petals of male inflorescences and pollen consumption. This is because female inflorescences have lignified cells (sclerenchyma) and acicular crystals (raphides) that may protect their structures against herbivory [50]. Protective attributes have been reported in inflorescences in different species of the dioecious palm Chamaedorea sp., where certain morphological features in the flowers including sclerified tissue, silica bodies and raphide-containing ideoblasts are represented differentially in the sexual phases of palms [68, 69]. The greatest protection against herbivory in female inflorescences is possibly associated with protection of the fruits that will develop later. Therefore, the largest richness of visitors on male inflorescences is associated with herbivorous species, which consume flowers and pollen of A. mexicanum mainly due to high nutritional content in terms of proteins, lipids and carbohydrates [70, 71]. In addition, other species are attracted to male flowers to prey on some arthropods visiting these flowers [50].
Based on previous studies, we know that mutualistic networks are significantly nested when evaluated at the community level (e.g., seed dispersal, pollination, and protective networks) [51, 72, 73]. However, the nested pattern has only recently been reported in other types of ecological networks at the intrapopulation level [28, 29, 31, 74]. Here we showed that nestedness is also a non-random pattern of individual-based palm-floral visitor networks. This indicates that the floral visitors of A. mexicanum did not forage opportunistically, since the less visited individuals (visited only by pollinators) are a subset of the more visited individuals (visited by both effective and non-effective pollinators). We hypothesize that variation between individual plants (e.g. floral rewards) associated with floral resources might be the main traits that generated nested patterns [27]. Moreover, we observed that within the populations studied, there are no groups of pollinator species visiting more frequently a particular group of individual plants and generating highly modular networks. This can occur because all floral visitors of a plant species must have the same ability to visit all the individuals of such species. Therefore, it is expected that pollination networks can be highly modular only when different plant species are evaluated, in which some plant species are adapted to particular groups of pollinators (i.e. the classical definition of pollination syndromes) [16, 75], and due to specialization to floral characteristics by the effective pollinators [76]. Additionally, it is possible that the sequential flowering of this palm’s phenology can also generate modular networks, once the first plants to flower could not be visited by the same visitors than more delayed flowering plants [77], which generate compartments with many intragroup links and few intergroup links.
We know that the few species found in the generalist core contribute considerably more to the nested pattern and to the stability of the network compared to peripheral species [23]. In this study, although the peripheral composition of floral visitors in the networks changed across fragments, the core of generalist species remained stable over all fragments studied. Moreover, these species of floral visitors found in the generalist core are certainly the effective pollinators of A. mexicanum, such as E. centralis, M. mexicanus, C. aberrans, and Mystrops sp. On the other hand, we did not find any effect of the fragment size on the structure of individual-based plant-pollinator networks, which does not corroborate our initial prediction. Therefore, it is possible that the effective pollinators are maintaining the structure of these networks stable along the fragmentation gradient, and local extinction of such species may lead to loss of populations of A. mexicanum and the collapse of the networks within these fragments.
In this study we showed that individual-based palm-floral visitor networks may be more stable to forest fragmentation when compared with plant-pollinator networks studied at the community level, e.g., [22]. In addition, we observed that the magnitude and variance of effective pollinators of male and female flowers of A. mexicanum increased with forest fragment size. In this case, small fragments where individuals are less visited by effective pollinators could exacerbate the amount of pollen and diversity of pollen donors transferred between flowers (less variance in the abundance of pollinators per flower). Also the density of reproductive individuals can lead to an increase of potential donors, and have a positive impact on pollination success, as showed for the Neotropical palm Geonoma epetiolata [78]. These fragmentation-related scenarios could have strong consequences on the maintenance of plant species if we consider breeding systems of plants, as well as size and mobility capacity of pollinators, with consequences on some attributes of performance of seedlings in forest remnants [79]. In sum, we showed that beyond the structure of interaction networks, intrinsic and biological factors of plant-pollinator interactions within such networks are crucial to our knowledge of the conservation and maintenance of pollination networks in tropical forests. Moreover, given the prevalence of fragmentation in tropical ecosystems, our results indicate that the current patterns of land use will have consequences on the underlying mechanisms of pollination in remnant forests.
Acknowledgments
The Los Tuxtlas-UNAM Biological Station provided all necessary facilities to conduct this study. We thank Jéssica Falcão, Dulce Rodríguez, Stephen C. Pratt and one anonymous referee for their valuable improvements to this article. WD was supported by CNPq grant 237339/2012-9.
Author Contributions
Conceived and designed the experiments: AA RD. Performed the experiments: AA. Analyzed the data: WD. Contributed reagents/materials/analysis tools: WD AA MQ RD. Wrote the paper: WD AA MQ RD.
References
- 1. Fahrig L. Effects of habitat fragmentation on biodiversity. Annu Rev Ecol Syst. 2003; 34: 487–515.
- 2. Dirzo R, Young HS, Galetti M, Ceballos G, Isaac NJ, Collen B. Defaunation in the Anthropocene. Science. 2014; 345: 401–406. pmid:25061202
- 3. Del-Claro K, Torezan-Silingardi HM. Insect-plant interactions: new pathways to a better comprehension of ecological communities in Neotropical savannas. Neotropic Entomol. 2009; 38: 159–164
- 4. Memmott J, Waser NM, Price MV. Tolerance of pollination networks to species extinctions. P Roy Soc Lond B. 2004; 271: 2605–2611.
- 5. Dyer LA, Walla TR, Greeney HF, Stireman JO III, Hazen RF. Diversity of interactions: a metric for studies of biodiversity. Biotropica 2010; 42: 281–289.
- 6. Mello MA, Bezerra EL, Machado IC. Functional roles of centridini oil bees and malpighiaceae oil flowers in biome-wide pollination networks. Biotropica 2013; 45: 45–53.
- 7. Aguilar R, Ashworth L, Galetto L, Aizen MA. Plant reproductive susceptibility to habitat fragmentation: review and synthesis through a meta‐analysis. Ecol Lett. 2006; 9: 968–980. pmid:16913941
- 8.
Valiente-Banuet A, Aizen MA, Alcántara JM, Arroyo J, Cocucci A, Galetti M, et al. Beyond species loss: the extinction of ecological interactions in a changing world. Func Ecol. 2015; In press.
- 9. Tylianakis JM, Laliberté E, Nielsen A, Bascompte J. Conservation of species interaction networks. Biol Conserv. 2010; 143: 2270–2279.
- 10. Rathcke BJ, Jules E. Habitat fragmentation and plant-pollinator interactions. Curr Sci. 1994; 65: 273–278.
- 11. Kearns CA, Inouye DW, Waser NM. Endangered mutualisms: the conservation of plant-pollinator interactions. Ann Rev Ecol Syst. 1998; 29: 83–112.
- 12. Aizen MA, Feinsinger P. Habitat fragmentation, native insect pollinators, and feral honeybees in Argentine Chaco Serrano. Ecol Appl. 1994; 4: 378–392.
- 13. Cunningham SA. Effects of habitat fragmentation on the reproductive ecology of four plant species in mallee woodland. Conserv Biol. 2000; 14: 758–768.
- 14. Ashworth L, Quesada M, Casas A, Aguilar R, Oyama K. Pollinator-dependent food production in Mexico. Biol Conserv. 2009; 142: 1050–1057.
- 15. Potts S, Biesmeijer JC, Kremen C, Neumann P, Schweiger O, Kunin WE. Global pollinator declines: trends, impacts and drivers. Trends Ecol Evol. 2010; 25: 345–353 pmid:20188434
- 16. Olesen JM, Bascompte J, Dupont YK, Jordano P. The modularity of pollination networks. Proc Natl Ac Sc (USA). 2007; 104: 19891–19896. pmid:18056808
- 17. Dupont YL, Padrón B, Olesen JM, Petanidou T. Spatio-temporal variation in the structure of pollination networks. Oikos. 2009; 118: 1261–1269.
- 18. Vázquez D, Blüthgen N, Cagnolo L, Chacoff NP. Uniting pattern and process in plant–animal mutualistic networks: a review. Ann Bot. 2009; 103: 1445–1457 pmid:19304996
- 19. Clemente MA, Lange D, Dáttilo W, Del-Claro K, Prezoto F. Social wasp-flower visiting guild interactions in less structurally complex habitats are more susceptible to local extinction. Sociobiology. 2013; 60: 337–344.
- 20. Morales CL, Aizen MA. Invasive mutualisms and the structure of plant-pollinator interactions in the temperate forest of north-west Patagonia, Argentina. J Ecol. 2006; 94: 171–180.
- 21. Aizen MA, Sabatino M, Tylianakis M. Specialization and rarity predict non-random loss of interactions from mutualist networks. Science. 2012; 335: 1486–1489. pmid:22442482
- 22. Spiesman BJ, Inouye BD. Habitat loss alters the architecture of plant–pollinator interaction networks. Ecology. 2013; 94: 2688–2696 pmid:24597216
- 23.
Bascompte J, Jordano P. Mutualistic Networks. Princeton University Press; 2013.
- 24. Tur C, Vigalondo B, Trøjelsgaard K, Olesen JM, Traveset A. Downscaling pollen–transport networks to the level of individuals. J Anim Ecol. 2014; 83: 306–317. pmid:24107193
- 25.
Kunin WE. Population size and density effects in pollination: pollinator foraging and plant reproductive success in experimental arrays of Brassica kaber. J Ecol. 1997; 225–234.
- 26. Gómez JM, Perfectti F, Jordano P. The functional consequences of mutualistic network architecture. PLoS ONE. 2011; 6: e16143. pmid:21283583
- 27. Gómez JM, Perfectti F. Fitness consequences of centrality in mutualistic individual-based networks. Proc R Soc Lond B Biol Sci. 2012; 279: 1754–1760.
- 28. Araujo MS, Martins EG, Cruz LD, Fernandes FR, Linhares AX, dos Reis SF, et al. Nested diets: a novel pattern of individual-level resource use. Oikos. 2010; 119: 81–88.
- 29. Pires MM, Guimaraes PR, Aráujo MS, Giaretta AA, Costa JCL, dos Reis SF. The nested assembly of individual-resource networks. J Anim Ecol. 2011; 80: 893–903.
- 30. Tinker MT, Guimarães PR, Novak M, Marquitti FMD, Bodkin JL, Staedler BG, et al. Structure and mechanism of diet specialisation: testing models of individual variation in resource use with sea otters. Ecol Lett. 2012; 15: 475–483. pmid:22414160
- 31. Cantor M, Pires MM, Longo GG, Guimaraes PR, Setz EZF. Individual variation in resource use by opossums leading to nested fruit consumption. Oikos. 2013; 122: 1085–1093.
- 32. Dáttilo W, Serio-Silva JC, Chapman C, Rico-Gray V. Highly nested diets in intrapopulation tropical monkey-resource food webs. Am J Primatol. 2014; 76: 670–678. pmid:24497349
- 33. Dáttilo W, Fagundes R, Gurka CAQ, Silva MSA, Vieira MCL, Izzo TJ, et al. Individual-based ant-plant networks: diurnal-nocturnal structure and species-area relationship. PLoS ONE. 2014b; 9: e99838. pmid:24918750
- 34. Henderson A. A review of pollination studies in the Palmae. Bot Rev. 1986; 52: 221–259.
- 35. Scariot A. Forest fragmentation effects on palm diversity in central Amazonia. J Ecol. 1999; 87: 66–76.
- 36. Vormisto J. Palms as rainforest resources: How evenly are they distributed in Peruvian Amazonian?. Biodiv Conserv. 2002; 11: 1025–1045.
- 37.
Govaerts J, Dransfield J. World checklist of palms. The Board of Trustees of the Royal Botanic Gardens, Kew; 2005.
- 38. Kreft H, Sommer JH, Barthlott W. The significance of geographic range size for spatial diversity patterns in neotropical palms. Ecography. 2006; 29: 21–30.
- 39. Aguirre A, Dirzo R. Effects of fragmentation on pollinator abundance and fruit set of an abundant understory palm in a Mexican tropical forest. Biol Conserv. 2008; 141: 375–384.
- 40. Siefke R, Bernal R. Floral biology and insect visitors of the understory palm Synechantus warscewiczianus at the Pacific Coast of Colombia. Palms. 2004; 48: 33–41.
- 41. Núñez-Avellaneda LA, Rojas-Robles R. Reproductive biology and pollination ecology of the milpesos palm Oenocarpus bataua in the Colombian Andes. Caldasia. 2008; 30: 101–125.
- 42. García-Aguirre MC, Álvarez R, Dirzo R, Ortiz MA, Eng MM. Delineation of biogeomorphic land units across a tropical natural and humanized terrain in Los Tuxtlas, Veracruz, México. Geomorphology. 2010; 121: 245–256.
- 43. Dirzo R, Garcia MC. Rates of deforestation in Los Tuxtlas, a neotropical area in southeast Mexico. Conserv Biol. 1992; 6: 84–90.
- 44. Mendoza E, Fay J, Dirzo R. A quantitative analysis of forest fragmentation in Los Tuxtlas, southeast Mexico: patterns and implications for conservation. Rev Chil Hist Nat. 2005; 78: 451–467.
- 45.
Ibarra-Manríquez G, Martínez-Ramos M, Dirzo R, Nuñez-Farfán J. La vegetación. In: Gonzalez-Soriano E, Dirzo R, Vogt RC, editors. Historia Natural de Los Tuxtlas. Universidad Nacional Autónoma de México; 1997. pp. 61–85.
- 46.
Quero HJ. Palmae. Fasciculo 81. Flora de Veracruz. Xalapa, México; 1994.
- 47. Bongers F, Popma J, Meave del Castillo J, Carabias J. Structure and floristic composition of the lowland rain forest of Los Tuxtlas, Mexico. Vegetatio. 1988; 74: 55–80.
- 48. Aguirre A, Guevara R, Dirzo R. Effects of forest fragmentation on assemblages of pollinators and floral visitors to male- and female-phase inflorescences of Astrocaryum mexicanum (Arecaceae) in a Mexican rain forest. J Trop Ecol. 2011; 27: 25–33.
- 49. Piñero D, Sarukhán J, González E. Estudios demográficos en plantas. Astrocaryum mexicanum Libm. I. Estructura de poblaciones. Bol Soc Bot Mex. 1977; 37: 60–118.
- 50. Búrquez A, Pedroza AL. Floral biology of a primary rain forest palm, Astrocaryum mexicanum Liebm. Bot J Linn Soc. 1987; 94: 407–419.
- 51. Bascompte J, Jordano P, Melián CJ. Olesen JM. The nested assembly of plant-animal mutualistic networks. Proc Natl Acad Sci USA. 2003; 100: 9383–87. pmid:12881488
- 52. Blüthgen N, Menzel F, Blüthgen N. Measuring specialization in species interaction networks. BMC Ecol. 2006; 6: 12–18. pmid:16965615
- 53. Blüthgen N, Menzel F, Hovestadt T, Fiala B, Blüthgen N. Specialization, constraints, and conflicting interests in mutualistic networks. Curr Biol. 2007; 17: 341–346. pmid:17275300
- 54. Bersier LF, Banasek-Richte C, Cattin MF. Quantitative descriptors of food–web matrices. Ecology. 2002; 83: 2394–2407.
- 55. Dáttilo W, Dyer L. Canopy openness enhances diversity of ant–plant interactions in the Brazilian Amazon rainforest. Biotropica. 2014; 46: 712–719.
- 56. Tylianakis JM, Tscharntke T, Lewis OT. Habitat modification alters the structure of tropical host–parasitoid food webs. Nature. 2007; 445: 202–205. pmid:17215842
- 57. Blüthgen N, Fründ J, Vázquez DP, Menze F. What do interaction network metrics tell us about specialization and biological traits? Ecology. 2008; 89: 3387–3399. pmid:19137945
- 58. Almeida-Neto M, Guimarães-Jr P, Guimarães PR, Loyola RD, Ulhrich W. A consistent metric for nestedness analysis in ecological systems: reconciling concept and measurement. Oikos. 2008; 117: 1227–1239
- 59. Guimarães-Jr P, Guimarães PR. Improving the analyses of nestedness for large sets of matrices. Environ Modell Softw. 2006; 21: 1512–1513.
- 60. Guimerà R, Amaral LAN. Functional cartography of complex metabolic networks. Nature. 2005; 433: 895–900. pmid:15729348
- 61. Guimerà R, Sales-Pardo M, Amaral LAN. Modularity from fluctuations in random graphs and complex networks. Phys Rev E. 2004; 70: 025101. pmid:15447530
- 62. Marquitti FMD, Guimaraes PR, Pires MM, Bittencourt LF. MODULAR: Software for the Autonomous Computation of Modularity in Large Network Sets. Ecography. 2014; 37: 221–224.
- 63. Dáttilo W, Izzo TJ, Vasconcelos HL, Rico-Gray V. Strength of the modular pattern in Amazonian symbiotic ant-plant networks. Arthropod-Plant Interact. 2013; 7: 455–461.
- 64. Dáttilo W. Different tolerances of symbiotic and nonsymbiotic ant-plant networks to species extinctions. Net Biol. 2012; 2: 127–138.
- 65. Dáttilo W, Guimarães PR, Izzo TJ. Spatial structure of ant–plant mutualistic networks. Oikos. 2013; 122: 1643–1648.
- 66.
R Development Core Team. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. Available: http://www.r-project.org. Accessed 21 April 2012.
- 67. Thien LB, Azuma H, Kawano S. New perspectives on the pollination biology of basal angiosperms. Int J Plant Sci. 2000; 161: S225–S235.
- 68. Barfod AS, Hagen M, Borchsenius F. Twenty-five years of progress in understanding pollination mechanisms in palms (Arecaceae). Ann Bot. 2011; 108: 1503–1516. pmid:21831852
- 69. Askgaard A, Stauffer F, Hodel DR, Barfod AS. Floral structure in the Neotropical palm genus Chamaedorea (Arecoideae, Arecaceae). An Jard Bot Mad. 2008; 65: 197–210.
- 70. Uhl NW, Moore HE Jr. Correlations of inflorescence, flower structure, and floral anatomy with pollination in some palms. Biotropica. 1977; 9: 170–190.
- 71.
Rudall PJ (2007) Anatomy of flowering plants. Cambridge University Press. 139 p.
- 72. Guimarães PR, Sazima C, dos Reis SF, Sazima I. The nested structure of marine cleaning symbiosis: is it like flowers and bees? Biol Lett. 2007; 3: 51–54. pmid:17443964
- 73. Dáttilo W, Marquitti FMD, Guimarães PR, Izzo TJ. The structure of ant-plant ecological networks: is abundance enough?. Ecology. 2014; 95: 475–485. pmid:24669740
- 74. Araujo MS, Guimaraes PR, Svanback R, Pinheiro A, Guimaraes P, dos Reis SF, et al. Network analysis reveals contrasting effects of intraspecific competitions on individual vs. population diets. Ecology. 2008; 89: 1981–1993. pmid:18705384
- 75.
Faegri K, Van Der Pijl L. The principles of pollination ecology. Pergamon Press; 1980.
- 76. Stebbins GL. Adaptive radiation of reproductive characteristics in angiosperms, I: pollination mechanisms. Annu Rev Ecol Syst. 1970; 1: 307–326.
- 77. Vilela AA, Torezan-Silingardi HM, Del-Claro K. Conditional outcomes in ant–plant–herbivore interactions influenced by sequential flowering. Flora. 2014; 209: 359–366.
- 78. Martén-Rodríguez S, Quesada M, Momen B. Effects of local plant density and incomplete dichogamy on the reproductive success of the rare neotropical palm Geonoma epetiolata. Biotropica. 2012; 44: 680–688.
- 79. Aguilar R, Ashworth L, Calviño A, Quesada M. What is left after sex in fragmented habitats? Assessing the quantity and quality of progeny in the endemic tree Prosopis caldenia (Fabaceae). Biol Conserv. 2012; 152: 81–89.