Assessing plant-pollinator relationships often employs a snapshot approach to describe the complexity and dynamic involving species interactions. However, this framework overlooks the nuanced changes in species composition, their interactions, and the underlying drivers of such variations. This is particularly evident on less explored temporal scales, such as the dynamic decision-making processes occurring within hours throughout the day. To address these gaps, in this study, we evaluated the temporal and environmental factors shaping the change of species and interactions (beta diversity) between bees and plants throughout the day in a coastal environment in Mexico. In general, we found that the changes in species composition of bees and plants were mainly associated with species turnover throughout the day, while the principal component of changes in interaction composition was interaction rewiring (reassembling of pairwise bee-plant interactions). This was mainly because a few species (6 of 47 bee species, and 5 of 35 plant species) with many interactions remain permanent most of the day, leading to rewiring being the most important component of beta diversity interaction. While environmental conditions such as temperature and humidity did not significantly drive the compositional dissimilarity of species and interactions, we observed that nearby time intervals have a similar composition of species and interactions. In conclusion, our study emphasizes the importance of considering shorter temporal dynamics in understanding species interactions during the day. These insights deepen our understanding of the intricate dynamics shaping plant-pollinator interactions, providing valuable implications for future studies focused on conservation and management strategies.

Background and Aims Flower-visitor interactions comprise a continuum of behaviours, from mutualistic partners to antagonistic visitors. Despite being relatively frequent in natural communities, florivory remains unexplored, especially when comprising abiotic factors, spatio-temporal variations and global environmental changes. Here, we addressed the variation of florivory driven by changes in elevation and temporal flower availability. We expect decreased floral resources as elevation increases - due to environmental constraints - which may affect plant-florivore interactions. Yet if floral resources decrease but florivores remain constant, then we may expect an increase in florivory with increasing elevation in the community.Methods The flowering phenology of plant individuals was recorded in the Neotropical campo rupestre vegetation, in southeastern Brazil. Damage by florivores was recorded in plots at elevations ranging from 823 to 1411 m using two response variables as a proxy for florivory: the proportion of attacked flowers per plant and the proportion of petal removal on single flowers.Key Results Flower attack increased with elevation and damage was intensified in species with longer flowering periods. Conversely, longer flowering periods resulted in higher levels of petal removal with decreasing elevation. The temporal availability of flowers affected florivory, with the proportion of attacked flowers being more intense when there were fewer flowered individuals in the community. Petal removal on single flowers was intensified in plots with a larger number of individuals flowering, and with more species co-flowering.Conclusions This study provides one of the broadest records of a commonly neglected interaction of insects feeding on floral structures, quantifying the combined effect of floral display and availability along an elevation gradient in a highly biodiverse mountaintop community. These findings contribute to filling the gap in the understanding of florivory dynamics, focusing on a tropical mountaintop scenario facing imminent environmental changes and excessive natural resource exploitation.

Climate change is expected to affect the morphological, physiological, and life-history traits of plants and animal pollinators due to more frequent extreme heat and other altered weather patterns. This systematic literature review evaluates the effects of climate change on plant and pollinator traits on a global scale to determine how species responses vary among Earth's ecosystems, climate variables, taxonomic groups, and organismal traits. We compiled studies conducted under natural or experimental conditions (excluding agricultural species) and analyzed species response patterns for each trait (advance vs. delay or no change for phenology, decrease vs. increase or no change for other traits). Climate change has advanced plant and animal phenologies across most Earth's biomes, but evidence for temporal plant-pollinator mismatches remains limited. Flower production and plant reproductive success showed diverse responses to warming and low water availability in Alpine and Temperate ecosystems, and a trend for increased or neutral responses in Arctic and Tropical biomes. Nectar rewards mainly experienced negative effects under warming and drought across Alpine and Temperate biomes, but scent emissions increased or changed in composition. Life form (woody vs. nonwoody species) did not significantly influence trait response patterns to climate change. Pollinator fecundity, size, life-history, developmental, and physiological traits mostly declined with warming across biomes; however, animal abundance and resource acquisition traits showed diverse responses. This review identified critical knowledge gaps that limit our understanding of the impacts of climate change, particularly in tropical/subtropical biomes and southern latitudes. It also highlights the urgent need to sample across a greater range of plant families and pollinator taxa (e.g., beetles, wasps, vertebrates). The diversity of climate change effects should be assessed in the context of other anthropogenic drivers of global change that threaten critically important pollination interactions.

Bees play a pivotal role as pollinators in crops essential for human consumption. However, the global decline in bee populations poses a significant threat to pollination services and food security worldwide. The loss and degradation of habitats due to land use change are primary factors contributing to bee declines, particularly in tropical forests facing high deforestation rates. Here, we evaluate the pollination services provided to crops of watermelon (Citrullus lanatus) and green tomato (Physalis ixocarpa) in three municipalities in the state of Jalisco, Mexico, a place with Tropical Dry Forest, during years 2008, and 2014 to 2017. Both crops are cultivated in the dry season, approximately during the months of November to March. We describe the composition of the pollinator community and their visitation frequency (measured through the number of visits per flower per hour), and we assess the impact of pollinators on plant reproductive success and the level of pollinator dependence for each crop species (measured through the number of flowers that developed into fruits). We also evaluate how the landscape configuration (through the percentage of forest cover and distance to the forest) influences richness and abundance of pollinators (measured as number of species and individuals of pollinators per line of 50 m), and we use the model Integrated Valuation of Ecosystem Services and Tradeoffs (InVEST) to map and value the pollination service in both crops. InVEST Crop pollination model is a simulation focuses on wild pollinators providing the pollinator ecosystem service. Our findings indicate that Apis mellifera was the primary pollinator of both crops, one of the few abundant pollinators in the study region during the dry season. In experiments where pollinators were excluded from flowers, watermelon yielded no fruits, while green tomato experienced a 65% reduction in production. In the case of green tomato, fruit set showed a positive correlation with pollinator abundance. A positive association between forest cover and total pollinator abundance was observed in green tomato in 2008, but not in watermelon. Additionally, a positive relationship was observed between the abundance of bees predicted by the InVEST model and the abundance of bees observed in green tomato flowers in 2008. In the study region, green tomato and watermelon rely on pollinators for fruit production, with honeybees (from feral and managed colonies) acting as the primary provider of pollination services for these crops. Consequently, the conservation of natural areas is crucial to provide food and nesting resources for pollinators. By doing so, we can ensure the diversity and abundance of pollinators, which in turn will help secure food security. The findings of this study underscore the critical need for the conservation of natural areas to support pollinator populations. Policymakers should prioritize the protection and restoration of habitats, particularly tropical forests, which are essential for maintaining the diversity and abundance of pollinators.

Natural fires represent a frequent and widespread disturbance in many environments around the world and consequently plays important roles on ecosystem dynamics. The goal of our study was to evaluate the short-term effects of fire on termites in a tropical mountain ecosystem of the Brazilian Cerrado (savanna). Termites were sampled in unburned and burned areas 2 weeks after an extensive fire event. Sampling method consisted of toilet paper baits placed at three elevations: 800 m (Cerrado woodland habitat), 1100 and 1400 m (campo rupestre vegetation type). A total of 19 termite species were sampled, of which 15 occurred at 800 m, eight at 1100 m, and two at 1400 m. Average termite species richness and bait colonization were lower at 1400 m compared with the other elevations, which did not differ from each other. No effect of fire on average species richness and bait colonization was found for any elevation. As termites are known to be strongly affected by climate, soil, and vegetation, termite assemblages are probably being filtered physiologically by environmental conditions along the elevation gradient. The high tolerance of termites to fire effects is likely associated with the strategy of many collected species in constructing hard clay mounds, either above or below ground, or taking shelter in mounds built by other termite species to escape fire. We conclude that while termites are sensitive to environmental changes in conditions along the elevation gradient, they are also highly tolerant to fire in this tropical mountain.

Most studies evaluating insect communities along climatic gradients in mountain regions have overlooked the relevance of the contrasting effects of xeric and mesic habitats on species distribution and richness. Mesic zones in xeric ecosystems can provide more favorable conditions for the maintenance of some insects, including orchid bees (Apidae, Euglossini), which are known for their high species diversity in humid tropical forests. For the first time, we sought to understand the distribution of Euglossini in xeric and mesic habitats during dry and wet seasons along an elevational gradient in Espinhaço mountains, southeastern Brazil. The study collected 4873 orchid bees from three genera and 14 species. During the wet season, bee richness decreased with elevation in both xeric and mesic habitats. However, during the dry season, there was an increase in both bee richness and abundance at intermediate elevations. In the dry season, lower bee richness and abundance were recorded in xeric habitats compared with mesic habitats. Bee richness and abundance did not differ between the dry and wet seasons for mesic habitats, while lower richness and abundance were recorded in the dry season for xeric habitats. Species composition remained unchanged across habitats, and species turnover was identified as the primary driver of β ‐diversity among elevations and seasons. Orchid bees require high temperatures, but they cannot tolerate water stress in xeric habitats during the dry season. This highlights the importance of mesic habitats for maintaining these pollinators, which are essential for ecosystems in savanna and mountainous regions.

Knowledge on the distribution of mosquito communities over time and across human-modified landscapes is important in determining the risk for vector-borne disease. The diversity of mosquitoes along a rainy season and edge effects were evaluated in a riparian forest in the Cerrado biome, Southeastern Brazil. Mosquito communities were sampled with Shannon traps in three distinct habitats (forest interior, forest edge and pasture) throughout an entire rainy season, comprising five sampling months (December 2015 to April 2016). A total of 13 549 mosquitoes belonging to 54 species were sampled. Mosquito species richness and abundance were greater in February, which coincided with the middle of the rainy season and just after the months with greater rainfall. Mosquito species richness did not differ among habitats for any particular month. In February, month when 74% of individuals were recorded, mosquito abundance was lower in the pasture compared with the forest edge and interior, which did not differ statistically from each other. Four of the six most abundant mosquito species (which account for 93.5% of the sampled individuals) had more individuals collected in the forest edge, and 28 species were more abundant at the edge compared with 15 species in the forest interior. Months with high rainfall probably allowed the availability and maintenance of high-water level in breeding sites leading to a further increase in mosquito populations. While the pasture did not seem to have the ideal abiotic conditions and/or resources (e.g. food and breeding sites) for mosquito species, edge effects appear to favour mosquito populations. Therefore, the risk of mosquito-borne diseases is expected to be greater in the middle of the rain season at the riparian forest-pasture edge, when and where a greater number of disease-vectoring species are present.

In nature, leaf rolls form during the expansion of young leaves of some plant species but their importance as shelter structures for arthropods is unknown. We examined the role of expanding rolled leaves of aroid species as shelter sites for arthropods and evaluated the effect of leaf size and environmental context (natural forest vs pastures) on the occupation of these leaves. In a Mexican tropical rainforest, arthropods were sampled from 25 fully expanded and 25 expanding rolled leaves of each of three aroid species ( Rhodospatha wendlandii, Dieffenbachia oerstedii, and Spathiphyllum ortgiesii ), that differ in the size of their leaves. In addition, for D. oerstedii we also collected leaves from a pasture area. We recorded 120 arthropods (44 in expanded and 76 in rolled leaves) from 53 unique morphospecies (21 and 38). Average arthropod abundance was greater in rolled leaves compared with expanded leaves. Average arthropod species richness and abundance, and colonization frequency were lower in the aroid species with smaller leaves ( R. wendlandii ). Although we found no significant effects of habitat type on arthropods in rolled leaves of D. oerstedii , 3.4 times more individuals (34) and 2.11 times more morphospecies (19) were found in the rolled leaves in the forest compared with those in pasture (10 and 9). This study demonstrates that expanding rolled leaves of aroids, despite being an ephemeral resource, represent suitable leaf shelters used by a diverse arthropod fauna, and that leaf size and environmental context are important factors that affect the colonization of these leaves by arthropods.

Interactions between plants and herbivores are central in most ecosystems, but their strength is highly variable. The amount of variability within a system is thought to influence most aspects of plant-herbivore biology, from ecological stability to plant defense evolution. Our understanding of what influences variability, however, is limited by sparse data. We collected standardized surveys of herbivory for 503 plant species at 790 sites across 116° of latitude. With these data, we show that within-population variability in herbivory increases with latitude, decreases with plant size, and is phylogenetically structured. Differences in the magnitude of variability are thus central to how plant-herbivore biology varies across macroscale gradients. We argue that increased focus on interaction variability will advance understanding of patterns of life on Earth.

The construction of shelters on plants by arthropods might influence other organisms via changes in colonization, community richness, species composition and functionality. Arthropods, including beetles, caterpillars, sawflies, spiders, and wasps often interact with host plants via the construction of shelters, building a variety of structures such as leaf ties, tents, rolls, and bags; leaf and stem galls, and hollowed out stems. Such constructs might have both an adaptive value in terms of protection (i.e., serve as shelters) but may also exert a strong influence on terrestrial community diversity in the engineered and neighboring hosts via colonization by secondary occupants. While different traits of the host plant (e.g., physical, chemical and architectural features) may affect the potential for ecosystem engineering by insects, such effects have been, to a certain degree, overlooked. Further analyses of how plant traits affect the occurrence of shelters may thus enrich our understanding of the organizing principles of plant-based communities. This dataset includes more than a thousand unique records of ecosystem engineering by arthropods, in the form of structures built on plants. All records have been published in the literature, and span both natural structures (90.6% of the records) and structures artificially created by researchers (9% of the records). The data were gathered between 1932 and 2021, across more than 50 countries and several ecosystems, ranging from polar to tropical zones. Besides data on host plants and engineers, we aggregated data on the type of constructs and the identity of inquilines using these structures. This dataset highlights the importance of these subtle structures for the organization of terrestrial arthropod communities, enabling hypotheses testing in ecological studies addressing ecosystem engineering and facilitation mediated by constructs.

The magnitude of facilitation by shelter-building engineers on community structure is expected to be greater when they increase limited resources in the environment. We evaluated the influence of local environmental context on the colonisation of leaf shelters by arthropods in a Mexican evergreen tropical rainforest. We compared the species richness and abundance of arthropods (total and for different guilds) colonising artificially rolled leaves in habitats differing in understory heterogeneity (forest edge > old-growth forests > living fences). Arthropod abundance of the most representative arthropod taxa (i.e., Araneae, Blattodea, Collembola and Psocoptera) colonising the rolled leaves was greater at forest edge, a trend also observed for average arthropod abundance, and for detritivore and predator guilds. In addition, fewer arthropod species and individuals colonised the rolled leaves in the living fence habitat, a trend also observed for most arthropod guilds. As forest edge is expected to have a greater arthropod diversity and stronger density-dependent interactions, a greater limitation of refuges from competitors or predators may have determined the higher colonisation of the rolled leaves in this habitat. Our results demonstrate that local environment context is an important factor that affects the colonisation of arthropods in leaf shelters.

The construction of shelters on plants by arthropods might influence other organisms via changes in colonization, community richness, species composition, and functionality. Arthropods, including beetles, caterpillars, sawflies, spiders, and wasps often interact with host plants via the construction of shelters, building a variety of structures such as leaf ties, tents, rolls, and bags; leaf and stem galls, and hollowed out stems. Such constructs might have both an adaptive value in terms of protection (i.e., serve as shelters) but may also exert a strong influence on terrestrial community diversity in the engineered and neighboring hosts via colonization by secondary occupants. Although different traits of the host plant (e.g., physical, chemical, and architectural features) may affect the potential for ecosystem engineering by insects, such effects have been, to a certain degree, overlooked. Further analyses of how plant traits affect the occurrence of shelters may therefore enrich our understanding of the organizing principles of plant-based communities. This data set includes more than 1000 unique records of ecosystem engineering by arthropods, in the form of structures built on plants. All records have been published in the literature, and span both natural structures (91% of the records) and structures artificially created by researchers (9% of the records). The data were gathered between 1932 and 2021, across more than 50 countries and several ecosystems, ranging from polar to tropical zones. In addition to data on host plants and engineers, we aggregated data on the type of constructs and the identity of inquilines using these structures. This data set highlights the importance of these subtle structures for the organization of terrestrial arthropod communities, enabling hypotheses testing in ecological studies addressing ecosystem engineering and facilitation mediated by constructs. There are no copyright restrictions and please cite this paper when using the data in publications.

Current climate change is disrupting biotic interactions and eroding biodiversity worldwide. However, species sensitive to aridity, high temperatures, and climate variability might find shelter in microclimatic refuges, such as leaf rolls built by arthropods. To explore how the importance of leaf shelters for terrestrial arthropods changes with latitude, elevation, and climate, we conducted a distributed experiment comparing arthropods in leaf rolls versus control leaves across 52 sites along an 11,790 km latitudinal gradient. We then probed the impact of short- versus long-term climatic impacts on roll use, by comparing the relative impact of conditions during the experiment versus average, baseline conditions at the site. Leaf shelters supported larger organisms and higher arthropod biomass and species diversity than non-rolled control leaves. However, the magnitude of the leaf rolls' effect differed between long- and short-term climate conditions, metrics (species richness, biomass, and body size), and trophic groups (predators vs. herbivores). The effect of leaf rolls on predator richness was influenced only by baseline climate, increasing in magnitude in regions experiencing increased long-term aridity, regardless of latitude, elevation, and weather during the experiment. This suggests that shelter use by predators may be innate, and thus, driven by natural selection. In contrast, the effect of leaf rolls on predator biomass and predator body size decreased with increasing temperature, and increased with increasing precipitation, respectively, during the experiment. The magnitude of shelter usage by herbivores increased with the abundance of predators and decreased with increasing temperature during the experiment. Taken together, these results highlight that leaf roll use may have both proximal and ultimate causes. Projected increases in climate variability and aridity are, therefore, likely to increase the importance of biotic refugia in mitigating the effects of climate change on species persistence.

Hemiparasitic plants can reduce photosynthesis and alter the host plant’s foliar traits. This relationship may be more intense in ecosystems with nutritionally poor soil, such as the campo rupestre, known for its high endemism. This work evaluated the impact of a generalist hemiparasitic species (Struthanthus flexicaulis) on the physiology as well as physical and chemical traits of a native species (Baccharis dracunculifolia) frequently found in the campo rupestre. We planted twenty 60-day-old B. dracunculifolia individuals in a campo rupestre area. Two year after planting, during the 2019 dry season, we evaluated the water potential (Ψw), photosynthetic efficiency (PE), specific leaf area (SLA) and succulence (SU) of parasitized and non-parasitized host plants. These parameters were measured again in the dry season of 2020, when we also measured chlorophyll, nitrogen balance index (NBI) and phenolic content of leaves of the same individuals. Parasitized individuals had a higher Ψw and PE compared to non-parasitized individuals during the most critical period of the day (12:00 pm to 3:00 pm). However, towards the end of the day, parasitized individuals had lower Ψw and PE than non-parasitized ones. There was no effect of parasitism on SLA and SU. Parasitized plants had higher NBI and lower phenolic content than non-parasitized plants. We concluded that S. flexicaulis parasitism interfered in the physiology and chemical leaf traits of B. dracunculifolia. Parasitism, while momentarily increasing the PE and Ψw of parasitized individuals, can cause physiological imbalances, which could negatively affect the development and lead to the death of the host plant over time.

Burrow builders are often classified as ecosystem engineers because their digging activities regulate the availability of resources for other organisms. As antlion larvae prefer to construct their traps in bare sandy or loose soil, they could benefit from burrowing activities. We investigated the role of burrow builders as ecosystem engineers for antlions (Myrmeleontidae) in a tropical semideciduous forest in Mexico. The number of traps of antlion larvae was recorded on 30 sampling quadrats (45 cm) at the entrance of burrows (of unidentified builders) and on 30 paired off-burrow quadrats. Additionally, the percentage of bare soil was estimated for the 60 quadrats sampled. Of the 30 quadrats at the entrance of burrows, a total of 336 traps were recorded, with 21 (70%) of them having at least one trap, while for the 30 off-burrows quadrats, only two (6.6%) of them had traps, just three in total. The percentage of bare soil and the abundance of traps were significantly greater in quadrats at the entrance of burrows compared to quadrats without burrows. The abundance of traps at the entrance of burrows was positively affected by the percentage of bare soil. The few traps in the off-burrows quadrats suggested that, in addition to the limited bare soil, ground compactness probably limits the establishment of antlion larvae. Otherwise, when digging, burrow builders create small patches of bare sandy soils that are used by these insects. We concluded that the ecosystem engineering effect of burrow builders is an important structuring element for antlion populations in the tropical semideciduous forest studied.

In ant-plant mutualistic interactions, plants provide shelter (domatia) and/or food to ants and in exchange, and ants provide protection against herbivores. After plant tissue senescence, ants are expected to abandon dead domatia, leaving these empty spaces available for other arthropods. In this study, we tested for the role of the mutualistic interaction between Cordia alliodora and Azteca pittieri in promoting new habitats for arthropods through abandoned dead domatia. We predicted that species richness, abundance, and colonization frequency of secondary arthropods would be greater in dead branch domatia (dead domatia) compared with live branch domatia (live domatia). During March 2019, we selected 38 C. alliodora trees in a Mexican tropical dry forest. For each tree, we collected five live and five dead domatia, for 380 domatia in total. We found six morphospecies of secondary arthropods colonizing live domatia, while 42 were present in dead domatia. Ants were the most species-rich group (10 species) in abandoned dead domatia and utilized them as nesting sites (25 nests). Secondary arthropod species richness, abundance, and colonization frequency were greater in dead domatia compared with live domatia. We concluded that the Azteca-Cordia mutualistic interaction is an important habitat facilitator by promoting new habitats for arthropods through abandoned dead domatia.

Current climate change is disrupting biotic interactions and eroding biodiversity worldwide. However, species sensitive to drought, high temperatures and climate variability might persist in microclimatic refuges, such as leaf shelters built by arthropods. We conducted a distributed experiment across an 11,790 km latitudinal gradient to explore how the importance of leaf shelters for terrestrial arthropods changes with latitude, elevation and underlying climate. Our analyses revealed leaf shelters to be key facilitative elements for the diversity of arthropods. Predator diversity and overall biomass within shelters increased with local drought and temperature variability, regardless of latitude and elevation. In contrast, shelter usage by herbivores increased with abundance of predators on those same plants and in wetter climates. Projected increase in climatic variability and drought in certain geographic regions is therefore likely to enhance the importance of biotic refuges, especially for predators, in mitigating the impact of climate change on species persistence.

The Program for Biodiversity Research (PPBio) is an innovative program designed to integrate all biodiversity research stakeholders. Operating since 2004, it has installed long-term ecological research sites throughout Brazil and its logic has been applied in some other southern-hemisphere countries. The program supports all aspects of research necessary to understand biodiversity and the processes that affect it. There are presently 161 sampling sites (see some of them at Supplementary Appendix), most of which use a standardized methodology that allows comparisons across biomes and through time. To date, there are about 1200 publications associated with PPBio that cover topics ranging from natural history to genetics and species distributions. Most of the field data and metadata are available through PPBio web sites or DataONE. Metadata is available for researchers that intend to explore the different faces of Brazilian biodiversity spatio-temporal variation, as well as for managers intending to improve conservation strategies. The Program also fostered, directly and indirectly, local technical capacity building, and supported the training of hundreds of undergraduate and graduate students. The main challenge is maintaining the long-term funding necessary to understand biodiversity patterns and processes under pressure from global environmental changes.

Global sustainability rests on a myriad of benefits provided by natural ecosystems that support human livelihoods and well-being, from biodiversity persistence to climate regulation. The undeniable importance of conserving tropical forests has drawn most of the conservation spotlight towards it. However, open ecosystems such as the Brazilian Campo Rupestre (rupestrian grassland), have been historically overlooked despite their high diversity and key associated ecosystem services. We highlight major current threats to the persistence of the Campo Rupestre emphasizing its ecological, social, cultural, geo-environmental, and economic importance. We call attention to the importance of the Campo Rupestre as a reservoir of biodiversity and ecosystem services and offer priority actions that resulted from discussions involving scientists, industry representatives, environmental managers, and other members of civil society. Proposed actions include efforts related to ecological restoration, sustainable ecotourism, protection of traditional ecological knowledge, identification of emerging research questions, and development of tailored public policies. Such issues are integrated into a framework that collectively represents a road map to safeguard the Campo Rupestre from further degradation and steer its historical overexploitation towards sustainable management. Safeguarding the future of non-forest biomes like this poses a challenge to current paradigms of nature conservation. By establishing priorities and guidelines, we propose an actionable plan, which we hope can support informed decision-making policy towards a sustainable use of the Campo Rupestre.

Understanding the factors that influence the composition and structure of bee communities in natural habitats is critical for conservation and restoration efforts, mainly in disturbed ecosystems that are widely used for agricultural crop production, such as tropical dry forests (TDF). The aim of this study was to evaluate the effects of tree species composition and vegetation structure on the composition, abundance and species richness of eusocial and non-eusocial wild bee assemblages. Bees (19.909 individuals of 96 species) were collected in fifteen plots in different secondary succession stages in a TDF in the state of Minas Gerais, Brazil. We found a positive relationship between tree community similarity and bee community similarity, for both eusocial and non-eusocial bees. Average tree height positively affected the abundance and species richness of eusocial bees, while the abundance of non-eusocial bees was negatively affected by tree species richness. Similar tree species composition and vegetation structure between plots at the same stage of secondary succession probably determined more similar bee communities than dissimilar tree communities. Most eusocial bee species are dependent on cavities in large trees to house their colonies, while most non-eusocial bees probably prefer open habitats because they provide a greater density and diversity of floral resources, and suitable areas for ground-nesting species. Since eusocial bees represented 94% of the individuals sampled, the conservation and restoration of mature forests is of primary importance in order to increase eusocial bee diversity, and the maintenance of these areas in the vicinity of agricultural systems is crucial to increase the ecosystem service of pollination.