Thermally sensitive tropical bees are key contributors to specialized pollination services

Thermally sensitive tropical bees are key contributors to specialized pollination services | Authorea try { document.documentElement.classList.add('js'); } catch (e) { } var _gaq = _gaq || []; _gaq.push(['_setAccount', 'G-8VDV14Y67G']); _gaq.push(['_trackPageview']); (function() { var ga = document.createElement('script'); ga.type = 'text/javascript'; ga.async = true; ga.src = ('https:' == document.location.protocol ? 'https://ssl' : 'http://www') + '.google-analytics.com/ga.js'; var s = document.getElementsByTagName('script')[0]; s.parentNode.insertBefore(ga, s); })(); Skip to main content Preprints Collections Wiley Open Research IET Open Research Ecological Society of Japan All Collections About About Authorea FAQs Contact Us Quick Search anywhere Search for preprint articles, keywords, etc. Search Search ADVANCED SEARCH SCROLL This is a preprint and has not been peer reviewed. Data may be preliminary. 26 May 2025 V1 Latest version Share on Thermally sensitive tropical bees are key contributors to specialized pollination services Authors : Xiaojian Chen , Panpan Zhang , Chunxi Liang , Qiaoyi Nong , Feng Cai , Michael Orr 0000-0002-9096-3008 , Akihiro Nakamura , Yan-qiong Peng 0000-0002-7453-9119 , and Cheng Wenda 0000-0002-2596-3028 [email protected] Authors Info & Affiliations https://doi.org/10.22541/au.174828981.19447297/v1 253 views 214 downloads Contents Abstract Information & Authors Metrics & Citations View Options References Figures Tables Media Share Abstract Predicting warming impacts on pollinators is key for understanding and preserving biodiversity and ecosystem services under climate change. In the tropics, where most species exist, insect pollination is essential for functioning forests. However, the vulnerability of tropical insect pollinators such as bees to climate change is understudied. Existing studies lack links between thermal physiological metrics to pollination services. We quantified physiological heat tolerance, behavioural thermoregulation abilities, and their links to species role played in flower-visitation network for a tropical native bee community across different seasons. Physiological tolerance from 708 individuals of 92 bee species revealed strong phylogenetic signals and Halictidae exhibited ~3 ℃ lower heat tolerance than other bees in the hot-dry season when heat stress is the highest. Further quantification of body temperatures revealed Halictidae also had weaker ability to maintain stable body temperatures. Analyzing extensively sampled networks showed specialization index of bees, which is associated with pollination efficacy, was negatively correlated with heat tolerance in the hot-dry season. Specialized Halictidae interactions are therefore potentially more vulnerable to temperature rises, potentially leading to increasingly dominant generalized interactions in the future under climate change. Title: Thermally sensitive tropical bees are key contributors to specialized pollination services Authors : Chen, Xiaojian 1 ; Zhang, Panpan 1 ; Liang, Chunxi 1 ; Nong, Qiaoyi 2,3 ; Cai, Feng M. 1 ; Orr, Michael C 4 .; Nakamura, Akihiro 2 ; Peng, Yanqiong 2 ; Wenda, Cheng 1 * Affiliations : 1 School of Ecology, Shenzhen Campus of Sun Yat-sen University, Guangdong, China. 2 CAS Key Laboratory of Tropical Ecology, Xishuangbanna Tropical Botanical Garden; Mengla, Yunnan, China. 3 School of Ecology and Environmental Science, Yunnan University; Kunming, Yunnan, China. 4 Entomologie, Staatliches Museum für Naturkunde Stuttgart; Stuttgart, Germany. Emails : Chen, Xiaojian, [email protected] ; Zhang, Panpan, [email protected] ; Liang, Chunxi, [email protected] ; Nong, Qiaoyi, [email protected] ; Cai, Feng M., [email protected] ; Orr, Michael C [email protected] ; Nakamura, Akihiro, [email protected] ; Peng, Yanqiong, [email protected] ; Wenda, Cheng, [email protected] ; Statement of Authorship : CW conceptualized the idea; CW, CX, ZP, LC & NQ collected the data; AN & PY provided equipment for the experiments; CX & FC did the analysis; CX & CW drafted the manuscript; All co-authors read through and edited the manuscript extensively. Data accessibility statement : Data and code supporting the findings of this study could be accessed from the following temporary link in Dryad: http://datadryad.org/share/lKd7XA7UMQDVDyOueH_Uo08St4c3Qi61rD_adQtIUvU. The dataset included heat tolerance data of bees, bee phylogenetic tree, field measured body temperature of bees and corresponding microclimate scale air temperature, bee-flower visitation network. Running title : Specialized Halictids vulnerable to warming Keywords : Heat tolerance; thermoregulatory ability; network specialization; native bee Type of article : Letter Words in the abstract : 150 Words in the main text : 4576 The number of references : 66 The number of figures : 3 The number of tables : 1 Mailing address of corresponding author : Wenda, Cheng, [email protected] ; Phone: +860755-23260286, School of Ecology, Sun Yat-sen University Shenzhen Campus, Guangdong, China, 518000. Abstract: Predicting warming impacts on pollinators is key for preserving biodiversity and ecosystem services under climate change. However, the vulnerability of tropical bees to climate change is understudied, lacking links between thermal traits to pollination services. We quantified heat tolerance, thermoregulatory abilities, and their links to network specializations in flower-visitation networks for a tropical bee community across seasons. Physiological tolerance from 708 individuals of 92 bee species revealed strong phylogenetic signals and Halictidae exhibited ~3 ℃ lower heat tolerance than other bees in the season when heat stress is the highest. Field-measured body temperatures revealed Halictidae also had weaker ability to maintain stable body temperatures. Network analysis showed specialization index of bees, which is associated with pollination efficacy, was negatively correlated with heat tolerance in the hot-dry season. Specialized Halictidae interactions are therefore likely more vulnerable to warming, potentially leading to increasingly dominant generalized interactions in the future under climate change. Introduction Predicting climate change impacts on ecosystem service providers is vital for preserving overall functionality (Scheffers & Pecl 2019). Bees are responsible for pollination of ~90% angiosperm plants worldwide (Ollerton et al. 2011), but are feared to be declining (Dicks et al. 2021). Previous studies from historical surveys have recorded and predicted severe climate change impacts for some bees (Kerr et al. 2015; Soroye et al. 2020), and pollination services (Zoller et al. 2023). However, our current understanding of bee population and communities is predominantly based on temperate regions and limited taxa (primarily honeybees, bumblebees) (Goulson et al. 2015; Kerr et al. 2015; Soroye et al. 2020); the status of most species and drivers of decline remains poorly known. Bees from warmer tropical regions are not only less studied, but also likely under greater heat stress. This is because air temperatures are often higher in the tropics and bees generate excess heat during flight between flowers (Glass et al. 2024). Although the richness of bees is lower in the tropics (Orr et al. 2021), the proportion of plants that rely on insect pollination are much higher (Ollerton et al. 2011). Understanding how native bees’ response to warming in the tropics is thus important for maintaining reproduction of tropical plants and related ecosystem functions and services. To achieve this goal, warming responses of bees need to be translated to pollination metrics to make predictions on ecosystem services. Pollinator network specialization has been commonly used to represent the effectiveness of pollination services, because specialized bees that ensure con-specific pollen transfer exhibit high pollination effectiveness (Brosi 2016). Linking network specialization metrics to climate change sensitive traits such as heat tolerance is thus crucial (Schleuning et al. 2020; Kazenel et al. 2024). However, existing community-level studies examining heat tolerance in bees are mostly from temperate regions and focused on population or richness responses (Hamblin et al. 2017; Kazenel et al. 2024), lacking synergistic approaches to make links across levels from ecophysiology to communities and pollination services. While physiological heat tolerance is important, thermoregulation can also provide important buffers (Johnson et al. 2022), yet studies examining community-level bee thermoregulation are generally lacking (Herrera et al. 2023; Herrera 2024). Drivers of thermal adaptation-related trait variation are also unclear (Diamond and Yilmaz 2018). Meanwhile, thermal approaches are even rarer in network analyses (Ratoni et al. 2024), making synthetic frameworks for community-level bee ecophysiology and network specialization imperative under climate change. For these reasons, we carried out our study in a well-sampled tropical native bee community in southwestern China, representing the largest studied tropical native bee community to date. We measured heat tolerance, quantified thermoregulatory abilities with field measured body temperatures, and sampled associated bee-flora visitation networks intensively in a standardized approach across multiple seasons. We tested three hypotheses that built on theories and findings from previous thermal biology studies. 1) Environment temperature can shape heat tolerance (González‐Tokman et al. 2020) and physiological heat tolerance of ectotherms is often correlated with evolutionary history and body size (Kellermann et al. 2012; Kazenel et al. 2024). We tested whether heat tolerance of tropical bees vary with seasonal temperature differences and associate with phylogenetic relatedness and body size. 2) Given that temperate bees regulate their body temperature at different heating rates and body size can affect such thermoregulatory ability (Herrera 2024), we examined whether thermoregulatory ability of different bees in the tropics also vary with phylogeny and body size. 3) Because activity of bees is known to be sensitive to air temperature variations (Willmer & Stone 2004), to maximize resource intake from flower visits, we tested whether bee species that access a wider variety of resources (generalist pollinators) are selected for activity during hotter periods to maximize their resource intake and hence exhibit higher heat tolerances (whether specialist flower visitors are less tolerant to heat). Accordingly, we provide an unparalleled, multi-faceted exploration of bee thermal ecology in the tropics. Materials and Methods Field site and network sampling We conducted our study in Xishuangbanna Tropical Botanical Garden, Chinese Academy of Sciences, Mengla, Xishuangbanna Dai Autonomous Prefecture, Yunnan Province (21°41’ N, 101°25’ E, 570m a.s.l.). The area has a tropical monsoon climate with an average annual precipitation of 1200-1700 mm and an average annual temperature of 21.4℃. In our study area, there are three distinct seasons in a year: the hot and dry season (hereafter hot-dry season) from March to May, the hot and wet season from June to October (hereafter wet season), and the cool and dry season from November to February (hereafter cool season) (Wenda et al. 2023; Figure S1 ). Although the climate is seasonal, the forests are largely evergreen. In our field site, we set up three transects 100m long in forest edges1.98 ± 0.74 km apart from each other. Despite some limited horticultural and non-native species, native plants dominated the vegetation along our transects. In 2023 from April to December, we sampled bee-flora visitation networks across three seasons (hot-dry, wet, and cool). In each season, we sampled each transect for four days, totalling within two continuous weeks for all three transects. During each sampling day, we walked the same transect three times (9:00, 11:00, 13:00), each walk lasted ( Figure S1 ). We recorded bee species and their abundances visiting different flowers. Since some bee species are difficult to identify in the field (e.g., subgenus Zonamegilla in Amegilla ), we also collected bees using insect nets for closer examination and later heat tolerance measurement. To minimize the impact of netting on flora visitation and local native bee populations, we collected a maximum of three individuals for one species during a single day and alternated our sampling order for each transect. During each season sampling, we also put a data logger (HOBO MX2202, Onset, USA) along each transect in shaded area to measure air temperature at 1-min interval. The permit to collect insect specimens was issued by the Center for Gardening and Horticulture, Xishuangbanna Tropical Botanical Garden, Chinese Academy of Sciences. We stored the bee specimens in the museum of Xishuangbanna Tropical Botanical Garden. CTmax measurements, thermal safety margins (TSM), and body size After capture, we put the bees in an ~16℃ icebox before transfer to the lab. Before heat experiments, the bees were allowed to recover at room temperature (~26 ℃) until they were able to walk naturally. We put the bees into small but well-ventilated transparent plastic boxes (50*50*40mm). We put a piece of water saturated tissue paper (~25*40mm) at the bottom of each box to control for biases from desiccation. During the experiments, we put the plastic boxes into a 5 L thermal chamber that could control both relative humidity and temperature (TL KC-80, Tianling Instrument Co., Ltd. Jiangsu, China). We set a ramping rate of 1.2 ℃/min at the RH of 90% from 26 ℃ to 60℃ . The ramping rate used by previous studies ranged from 0.5 ℃/min ( Gonzalez et al. 2024 ) to 2 ℃/min (Ratoni et al. 2024). Faster ramping rates can overestimate CTmax, however, ramping rate will not affect our result interpretation as CTmax is a relative metric and are comparable as long as they were measured in a standardized way ( Rezende et al. 2011 ). Further, we choose this moderately fast ramping rate to shorten the experiment time and thus avoid bias caused by e nergy and oxygen exhaustion, to which bees are known to be susceptible due to their high metabolic rates (Heinrich & Esch 1994). We calculated the absolute difference between mean air temperature in each season obtained from our datalogger we mentioned earlier and CTmax of each bee species in corresponding seasons as their thermal safety margin (TSM). L ower TSM values indicate higher vulnerability under warming. TSM has many forms and as with CTmax, it is also a relative concept, and the absolute value has limited ecological inferences (i.e., narrower TSM indicated comparatively higher vulnerability but not necessarily imminent threats under warming) (Clusella-Trullas et al. 2021) . After measuring CTmax, we made bee specimens for further identification and barcoding (see the next section for detail). We measured the commonly used inter-tegular distance (ITD) for each bee specimen as proxy for body size ( Cane 1987 ). not-yet-known not-yet-known not-yet-known unknown Bee identification and phylogenetic tree reconstruction Species-level identification is challenging at our site, because many species were described based on only one sex yet sexual dimorphism is common (especially for Megachilidae). For these reasons, we first applied a morpho-species approach and then used DNA barcoding to verify the validity of our morpho-species and associate males to their corresponding females. Our criteria for morpho-species relied largely on morphological differences distinguishable in the field, so that we could release the bees after checking to minimize the impact to the network. Barcode results suggest that these were aligned with true species, which is often feasible for less diverse tropical faunas (Orr et al. 2021). Further, we also dissected the male genitalia of difficult groups to confirm their distinctness and compare with published literature to improve our understanding of local bee diversity. For DNA barcoding, we used leg tissue from our bee specimens to extract the total DNA that was then used for partially amplifying the mitochondrial Cytochrome C oxidase subunit I (COI) gene for sequencing. The primers of C1-J-2183 (alias Jerry) were adopted for this purpose following the PCR protocol given in Simon et al. (1994). We used the obtained COI sequences to construct the phylogenetic tree. We generated the multiple sequences alignment (MSA) by using MUSCLE (version 3.8.31, Edgar, 2004), and we carried out a maximum-likelihood (ML) phylogenetic analysis using IQ-TREE 1.6.12 (Nguyen et al., 2015) based on the MSA, with 1000 ultrafast bootstrapping replicates. We searched for the best nucleotide substitution model using ModelFinder (Kalyaanamoorthy et al., 2017) according to the Bayesian Information Criterion (BIC). We visualized generated phylogenetic tree in iTOL (https://itol.embl.de/). Any individuals that we were unsure of their identity but clustered with other known morpho-species were considered the same species. not-yet-known not-yet-known not-yet-known unknown Combining morphological differences with DNA barcoding, we established reliable morphological differences for several species previously difficult to identify. For example, some of the most challenging groups were the Amegilla subgenus Zonamegilla and the species-rich genus Lasioglossum. In addition, we could also associate females with males in the genus Megachile for several species. Although we could only separate the cleptoparasitic genus Thyreus reliably based on dissecting the male genitalia, they are generally rare, and only less than 1% of our dataset were discarded due to taxonomy uncertainty. not-yet-known not-yet-known not-yet-known unknown Thermoregulatory ability assessment In 2024 from May to December, we sampled the Tb of lower visiting bees with an infrared thermal image camera (±2 °C, Fluke Ti480 Pro, USA) and simultaneously logged the fine scale Ta every minute with a shaded data logger (HOBO MX2202, Onset, USA) within 100m of the bees. The camera can focus on a minimum distance of 15cm with a resolution of 0.07mm at the distance of 15cm (0.14mm at the distance of 30cm). To increase the accuracy of the thermal image camera, we controlled the essential factors recommended by Playà‐Montmany & Tattersall (2021). We manually adjusted the focus to get clear shots of bees at the distance 15-20cm (for larger carpenter bees, our maximum focus distance is 30cm). We selected 0.97 as the emissivity of different bee species. This value was originally calibrated by Stabentheiner & Schmaranzer (1987) and has been used in other Hymenoptera (de Farias-Silva & Freitas 2021; Johnson et al. 2022). Since thorax temperature is important for flight activity of insects including bees (Heinrich & Esch 2004) and bees do prioritize their thorax temperature from our infrared photos, we extracted the maximum temperature of bee thorax within each photo as a Tb reading with the software Fluke Connect (Version 2.0.1.0, Fluke Corporation). To increase Tb estimation accuracy, we only included the thermal images that were in focus. We also only used species with at least five recordings over no less than 5°C air temperature range to increase the representation of thermoregulatory ability. We estimated thermoregulatory ability for each bee species with \(R_{\text{thermal}}\)(E.q. 1) from (Hertz et al. 1993). \[R_{\text{thermal}}=1-abs\left(\frac{T_{b}-T_{\text{pref}}}{T_{a}-T_{\text{pref}}}\right)\ (E.q.1)\] \(R_{\text{thermal}}\) represents how closely a bee’s body temperature (\(T_{b}\)) deviates from its preferred body temperature (\(T_{\text{pref}}\)) during activity relative to the available environmental temperature (\(T_{a}\)). \(R_{\text{thermal}}\) aligns with the principle that thermoregulation needs to account for deviation of body temperatures from their preferred body temperatures against the variation of the thermal environments (Hertz et al. 1993). For this reason, we did not use the metric \(T_{b}/T_{a}\) (Bishop & Armbruster 1999; Bladon et al. 2020). Because \(T_{\text{pref}}\) needs to be estimated under temperature gradient provided in controlled lab environments and it is difficult to estimate this for flying insects (Hertz et al. 1993), we used species mean body temperature recorded in the field (\(T_{b\_mean}\)) to proximate preferred \(T_{\text{pref}}\), assuming a normally distributed body temperature range would have the mean mostly favour by. The higher \(R_{\text{thermal}}\) is, the better a bee species can retain a stable body temperature against unfavourable air temperatures, indicating higher thermoregulatory ability. Specialization indices of bees We built quantitative bee-flower visitation networks by pooling the visitation data of all sampling days all transects within a same season together as our seasonal networks. Then, we calculated sampling completeness (Chao1 index) for each seasonal network (Chao 1984). We quantified bee specialisation in flora-visitation networks with DSI* using the R package dizzy (Jorge et al. 2017). DSI* accounted for the phylogenetic relationships among resource species (plants here) and is considered to represent resource use specialization more accurately. Higher values (towards 1) indicate higher specialisation. Because many species occurred in multiple seasons with seasonal CTmax variation, we calculated DSI* separately for each season. We used the R package U.PhyloMaker (Jin & Qian 2019) to generate phylogenetic trees of plants from a megatree base on the botanical nomenclature of the World Plants database (http://www.worldplants.de). Statistical analysis To test the contribution of evolutionary history to our main tested response variables (CTmax, TSM, \(R_{\text{thermal}}\), and DSI* ) , we separately calculated Pagel’s λ (Pagel 1999) with the R package caper (Orme et al. 2013). λ use Brownian motion to measure phylogenetic signal in traits, values close to 1 indicated strong phylogenetic signal and close to 0 indicated no signal. In general, we built different generalized linear models based on phylogenetic signal tests to examine our three questions. Firstly, we built phylogenetic generalized linear mixed models (PGLMM) with the R package phyr (Li et al. 2020) to examine co-linearity among body size, season, and family before fitting CTmax and TSM models. We selected the best models with Akaike’s information criterion (AIC) via the lowest AIC values. Since the best model showed no correlation among the three variables ( Table S2, S3 ), for the overall CTmax, CTmax in each season, and TSM in each season that have strong phylogenetic signal ( Table S1 ), we further used PGLMM to explore how these metrics was affected by body size, season, and family. We used these variables as fixed factors, and species and phylogeny as random factors. We did not include three-way interaction for the difficulty in interpretating the results. B oth CTmax and TSM models, we detected significant season effect as well as interactions among body size and bee family, but no independent body size and family effects. For this reason, we further examined the effects of body size and bee family in each season separately ( Table 1 ). For overall TSM, DSI*, and \(R_{\text{thermal}}\) that did not have significant phylogenetic signal ( Table S1 ), we used R package glmmTMB (Brooks et al. 2017) and built generalized linear mixed models (GLMM) with bee species as a random factor or generalized linear models (GLM) (if random factor is not in the best model) to examine how these variables responded to change of body size, season, and family ( Table S2 ). All model selection were made by the lowest AIC values. We did all analysis in R 4.4.2 (R Core Team 2023). Table 1. Summary of regression models used in different analyses. Because models for TSM use the same structure with CTmax apart from that the latter includes a phylogeny random factor, only models for CTmax are shown here. Detailed full model list, model selection, and statistical results are in Table S2 & S3 . Whether heat tolerance vary with season, family, and body size. CTmax ~ 1 + body size + season + family + family: body size + family:season + body size: season + (1|species) + phylogeny Individual level (n=708) In each season, whether heat tolerance vary with family and body size. CTmax ~ 1 + body size + family + family:body size + (1|species) + phylogeny Individual level (n= 299) Hot-dry season Individual level (n=284) Wet season Individual level (n=115) Cool season Whether thermoregulatory ability vary with taxa and body size Thermoregulatory ability ~ 1 + body size + family + family:body size + (1|species) Individual level (n=539) In each season, whether specialist flower visitors are less tolerant to heat DSI* ~ 1 + TSM + body size + (1|species) Species level (n=41) Hot-dry season Species level (n=51) Wet season Species level (n=22) Cool season Data summary Our three-season sampling obtained 3840 bee-flower visitation interactions. We measured 708 bee individuals for heat tolerance (CTmax) from 92 species. This included 26 bee species in the cool season, 54 species in the hot-dry season, and 61 species in the wet season ( Table S4 ). We sampled field measured body temperature (T b ) from 539 recordings of flower visiting bees, and quantified thermoregulatory ability (\(R_{\text{thermal}}\)) for 27 bee species. We measured body size of 614 individuals, ranging from 0.76 mm to 15.04 mm in ITD ( Figure S3 ). Mean air temperature was the highest during the wet season (26.8 ℃), followed by 26.0 ℃ during the hot-dry season and 21.3 ℃ during the cool season. Phylogenetic signals in thermal biology related traits Using our bee phylogeny, we identified strong phylogenetic signals for CTmax ( Fig 1a ; λ=0.71, 95% CI (0.49, 0.87)), and ITD (λ=1, 95% CI (0.96, 1)), but not for thermal safety margin (TSM) (λ=0.47, 95% CI (0.26, 0.69), \(R_{\text{thermal}}\) (λ=0, 95% CI (0, 0.67), nor network specialization index (DSI*) ( Table S1) . CTmax and TSM variation Although our best model for CTmax included all variables tested, we did not detect significant independent effect of body size and family, only season significantly affected CTmax of bees, with the wet season season and the hot-dry season, respectively ( Fig. 1b; Table S2, S3 ). TSMs of bees had similar pattern with no independent body size and family effect, while the TSM were ~ 3.1 ℃ and ~4.7 ℃ wider in the cool season compared to the hot-dry and wet seasons ( Fig. 1c ; Table S2, S3 ). CTmax analysis by season showed in both the hot-dry season and the cool season, family was significant and Halictidae had the lowest CTmax ( Table S2, S3 ). CTmax in Halictidae was ~3.5 ℃ lower than Apidae and Megachilidae in the hot-dry season and ( Table S2, S3 ). While in the wet season, although CTmax was also ~3.0 ℃ lower in Halictidae than other families, the difference was not significant, and CTmax was positively correlated with body size ( Table S2, S3 ). Since we used mean air temperature in each season to calculate TSM, the statistical results for TSM regression by season are the same with CTmax results in each season, and Halictidae exhibited narrower TSM compared with other families in the hot-dry and cool season, while TSM was positively correlated with body size in the wet season. Figure 1 . Patterns of heat tolerance and thermal safety margins by family, season, and phylogeny. (a), CTmax variation along phylogenetic relationships, seasonal means for each species are shown, darker red in the outer circle indicating higher CTmax. (b), CTmax variation across seasons and families, different alphabet indicated significance(p<0.05). (c), TSM variation across seasons and families, different alphabet indicated significance(p<0.05). Across three panels, different colours indicate different bee families as indicated in (c). Thermoregulatory ability Our PGLS found smaller bees have lower \(R_{\text{thermal}}\) and comparatively poorer ability to maintain preferred body temperatures (χ 2 =5.83, p=0.016; Fig. 2a; Table S2, S3 ). Megachilidae bees showed a relative higher \(R_{\text{thermal}}\) than Apidae and Halictidae bees ( Fig. 2b; Table S2, S3) . Figure 2. Patterns of behavioural thermoregulatory ability (\(R_{\text{thermal}}\)) of bees with different body sizes (a) and families (b). Different colours indicate different bee families. In (a), dashed vertical line indicates the body size (ITD: inter-tegular distance) median. Shaded area around the best fitted model (blue line) indicates standardized error (SE). In (b), different alphabet indicated significance(p<0.05). Correlation between network specialization and climate change vulnerability We found TSM was positively correlated with DSI* in the hot-dry season (χ 2 =4.98, p=0.02). However, for the cool season and the wet season, the correlation was insignificant ( Fig. 2; Table S2, S3 ). Figure 3 . Bee-flower visitation networks in each season (a, c, e) and corresponding correlation of species level specialization index (DSI*) and climate change vulnerability (TSM) (b, d, f). In (a), (c), and (e), the breadth of each colour band indicates interaction frequency, grey colour indicates shared interactions among three seasons, while red indicates unique interactions in each season, yellow indicates shared interactions between the hot-dry and wet season, cyan indicates interactions shared between the wet and cool season, blue indicates interactions share between the hot-dry and cool season. Each point in (b), (d), and (f) represents one species and different colours indicate different bee families. only significant correlation is labelled with fitted relationships. Shaded area around the best fitted model (black line) indicates standardized error (SE). Discussion Heat tolerance variation across taxonomic groups, seasons, and body sizes. Our study reveals heat tolerance variation of major tropical bee lineages in the region is strongly driven by evolutionary history, contrasting with results from temperate bees that find no phylogenetic signal in heat tolerance (Hamblin et al. 2017; Kazenel et al. 2024). Although our findings that Halictidae are less tolerant to heat across different seasons is unseen in temperate regions, this taxa-related climate change vulnerability has been shown in temperate bumblebees ( Bombus spp.) (Martinet et al. 2021). Possible explanations for why Halictidae have lower heat tolerance include that Halictidae has been found to be weaker thermoregulators in our study and in temperate regions (Herrera 2024). We show strong seasonal effects in heat tolerance, such that tropical bees from hotter seasons (the wet season) have higher heat tolerances, indicating thermal environments play important roles in shaping heat tolerance as with other bees (Martinet et al. 2021; Gonzalez et al. 2023). However, since the air temperature in the wet season is also the highest, the season with the narrowest thermal safety margins is the hot-dry season. In addition, our high seasonal variation of heat tolerance and high seasonal bee community similarity (Bray-Curtis: 0.74±0.04) indicates that thermal plasticity occurs within species (Seebacher et al. 2015). However, our strong phylogenetic signal in heat tolerance suggests evolutionary history may constrain such plasticity to a certain threshold (Gonzalez et al. 2024). We found that body size is important in affecting heat tolerance, but only in the wet season. Although body size is usually a key trait in affecting thermal adaptation in ectotherms (Peralta-Maraver & Rezende 2021), the effect of body size on heat tolerance for bee is in general mixed (Hamblin et al. 2017; Feuerborn et al. 2023), possibly due to bees being more endothermic than typical ectotherms (Heinrich & Esch 1994) and the complex roles body size play in bee thermal biology (Stone & Willmer 1989; Herrera 2024). Thermoregulatory ability variation across taxonomic groups and body sizes Although we detected no significant phylogenetic signal in thermoregulatory ability, we found Halictidae had lower thermoregulatory abilities compared to Megachilidae. This parallels a temperate study that found comparatively faster heat exchange rates with the environment and lower body temperatures maintained in Halictidae compared with Apidae and Megachilidae (Herrera 2024). Our weaker thermoregulatory ability of small body-sized tropical bees is also consistent with a temperate study that found smaller bees in general maintain less stable body temperatures (Bishop & Armbruster 1999). However, our results contrast with the ecophysiology theory “Bogert effect” in ectotherms that predicts effective behaviour thermoregulation can select conservative physiological tolerance (Bogert 1949; Muñoz 2022), and accordingly weaker thermoregulator (Halictidae here) should have higher heat tolerances. Ectotherms selecting microclimate to thermoregulate and Bogert effect may not rise if microclimates lack variability (Bodensteiner et al. 2021). It is likely our contrasting results are due to highly active bees maintaining higher thorax temperatures than air temperatures, making them less dependent on ambient temperatures (Heinrich & Esch 1994). However, other possible mechanisms explaining no Bogert effects include that Halictidae may have evolutionary constraints in heat tolerance (Muñoz 2022). Further experiments on thermoregulation behaviour and macroevolutionary studies focusing on Halictidae thermal adaptation is needed. Network specialization correlated with climate change vulnerability Our work establishes a negative correlation between climate change vulnerability and specialization of flower visitation in a tropical bee community, confirming previous predictions and observations that specialized links in the networks are more prone to disturbance and pollination networks likely become less efficient and more generalized in temperate ecosystems (Aizen et al. 2012; Miller-Struttmann et al. 2015; Bhandary et al. 2023; Zoller et al. 2023). Our negative correlation between climate change vulnerability and specialization of flower visitation is likely driven by the fact that most of the specialized flower visitors in the hot-dry season were halictids, which have been shown earlier to be less tolerant to heat in this season. Our results contradict a recent study that found bee species interacted with more plants were less tolerant to heat from coastal shrubs and forests (Ratoni et al. 2024). We suspect this difference may rise from ecosystems differences or methodology in quantifying specialization; for the latter, plant phylogeny is important and needs to be considered (Jorge et al. 2017). Further investigation on the mechanisms behind such differences is needed. High climate change vulnerability of Halictidae and small bodied bees in terms of pollination service We reveal that tropical Halictidae have weaker thermoregulatory abilities with lower heat tolerance and are highly susceptible to climate change. Halictidae are the second most diverse and abundant group of bees after Apidae (Michener, 2007). Most of Halictidae can buzz and are important for many tropical poricidal plants (Russell et al. 2024), and the family compromises 26 % of bee richness in our site. A diminished halictid community could have profound impacts on pollination in regional tropical ecosystems. Larger bees can deposit more pollen and often considered more effective in pollination (Földesi et al. 2021). Our findings of higher climate change vulnerability in smaller tropical bees in the wet season thus indicate a less pessimistic future for tropical pollination. However, body size is predicted and recorded to shrink under warming (Verberk et al. 2021), and existing monitoring from temperate ecosystems often found large-bodied bees showed greater population declines (Pardee et al. 2022). These evidences suggest multiple factors other than temperature and heat tolerance will influence bee survival. Long-term bee monitoring is needed in the tropics to confirm our prediction with body sizes and associated pollination service. Limitations and conclusion Although our thermal biology related traits can be used as proxies for warming responses, they do not necessarily represent species responses under climate change. We need to be cautious that other unmeasured traits, such as desiccation tolerance, may also be important (Kazenel et al. 2024). In addition, our study only covers one location in one tropical ecosystem and tropical ecosystems may function differently. However, our phylogenetically and taxonomically diverse bee community covers the major lineages of bees in the region (Ascher & Pickering 2020), and the seasonal tropical rainforest in our site also represents a widespread habitat in mainland southern Asia. Together, we reveal that tropical mutualist relationships are under disproportional threat from heat stress from physiological, behavioural, and ecology aspects, providing a robust evaluation of functionally important ectotherms and associated ecosystem functions under climate change. Specialized bees, particularly Halictidae that have weaker thermoregulatory ability, are under greater threat in the hot-dry season, when the highest daily maximum temperatures occur and pollinator richness is also the highest (Corlett 2019). These characteristics and our results highlight the significant roles hot-dry season plays in determining the survival of tropical native bees and pollination effectiveness. Our multi-approached extensive study provides a robust evaluation of functionally important ectotherms and associated ecosystem functions under climate change. We emphasize the importance of prioritizing the research and conservation of tropical Halictidae bees to better understand the fate of tropical bees and associated pollination services. Acknowledgements: We thank the Centre for Gardening and Horticulture, Xishuangbanna Tropical Botanical Garden, Chinese Academy of Sciences for permits. We thank Xia Xue, Miao Baige, & Dr. Xing Shuang’s for their assistance in the field logistics. We are also grateful for the comments from Dr. Louise A. Ashton and Dr. Timothy C. Bonebrake on the manuscript. Conflict of interest : The authors declare no competing interests. not-yet-known not-yet-known not-yet-known unknown References: Abou-Shaara, H. F., A. A. Owayss, Y. Y. Ibrahim, and N. K. Basuny. 2017. “A Review of Impacts of Temperature and Relative Humidity on Various Activities of Honey Bees.” Insectes Sociaux 64 (4): 455–63. Aizen, M. A., Sabatino, M., & Tylianakis, J. M. (2012). Specialization and rarity predict non-random loss of interactions from mutualist networks. Science, 335(6075), 1486-1489. Ascher, J. S., & Pickering, J. (2020, November). Discover Life bee species guide and world checklist (Hymenoptera: Apoidea: Anthophila). Bhandary, S., Deb, S., & Sharathi Dutta, P. (2023). Rising temperature drives tipping points in mutualistic networks. Royal Society Open Science, 10(2), 221363. Bishop, J. A., & Armbruster, W. S. (1999). Thermoregulatory abilities of Alaskan bees: effects of size, phylogeny and ecology. Functional Ecology, 13(5), 711-724. Bladon, A. J., Lewis, M., Bladon, E. K., Buckton, S. J., Corbett, S., Ewing, S. R., … & Turner, E. C. (2020). How butterflies keep their cool: Physical and ecological traits influence thermoregulatory ability and population trends. Journal of Animal Ecology, 89(11), 2440-2450. Bodensteiner, B. L., Agudelo‐Cantero, G. A., Arietta, A. A., Gunderson, A. R., Muñoz, M. M., Refsnider, J. M., & Gangloff, E. J. (2021). Thermal adaptation revisited: How conserved are thermal traits of reptiles and amphibians?. Journal of Experimental Zoology Part A: Ecological and Integrative Physiology, 335(1), 173-194. Bogert, C. M. (1949). Thermoregulation in reptiles, a factor in evolution. Evolution, 3(3), 195-211. Brooks, M. E., Kristensen, K., Van Benthem, K. J., Magnusson, A., Berg, C. W., Nielsen, A., … & Bolker, B. M. (2017). glmmTMB balances speed and flexibility among packages for zero-inflated generalized linear mixed modeling. The R journal, 9(2), 378-400. Brosi, B. J. (2016). Pollinator specialization: from the individual to the community. New Phytologist, 210(4), 1190-1194. Cane, J. H. (1987). Estimation of bee size using intertegular span (Apoidea). Journal of the Kansas Entomological Society, 145-147. Chao, A. (1984). Nonparametric estimation of the number of classes in a population. Scandinavian Journal of statistics, 265-270.Clusella-Trullas, S., Garcia, R. A., Terblanche, J. S., & Hoffmann, A. A. (2021). How useful are thermal vulnerability indices?. Trends in Ecology & Evolution, 36(11), 1000-1010. Clusella-Trullas, S., Garcia, R. A., Terblanche, J. S., & Hoffmann, A. A. (2021). How useful are thermal vulnerability indices?. Trends in Ecology & Evolution, 36(11), 1000-1010. Corlett, R. (2019). The ecology of tropical East Asia. Oxford University Press. de Farias-Silva, F. J., & Freitas, B. M. (2021). Thermoregulation in the large carpenter bee Xylocopa frontalis in the face of climate change in the Neotropics. Apidologie, 52(2), 341-357. Diamond, Sarah E, and Aaron R Yilmaz. 2018. “The Role of Tolerance Variation in Vulnerability Forecasting of Insects.” Current Opinion in Insect Science 29 (October):85–92. Dicks, L. V., Breeze, T. D., Ngo, H. T., Senapathi, D., An, J., Aizen, M. A., … & Potts, S. G. (2021). A global-scale expert assessment of drivers and risks associated with pollinator decline. Nature Ecology & Evolution, 5(10), 1453-1461. Edgar, RC. MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res. 2004, 32: 1792−1797. Feuerborn, C., Quinlan, G., Shippee, R., Strausser, T. L., Terranova, T., Grozinger, C. M., & Hines, H. M. (2023). Variance in heat tolerance in bumble bees correlates with species geographic range and is associated with several environmental and biological factors. Ecology and evolution, 13(11), e10730. Földesi, R., Howlett, B. G., Grass, I., & Batáry, P. (2021). Larger pollinators deposit more pollen on stigmas across multiple plant species—a meta‐analysis. Journal of Applied Ecology, 58(4), 699-707. Glass, J. R., Burnett, N. P., Combes, S. A., Weisman, E., Helbling, A., & Harrison, J. F. (2024). Flying, nectar-loaded honey bees conserve water and improve heat tolerance by reducing wingbeat frequency and metabolic heat production. Proceedings of the National Academy of Sciences, 121(4), e2311025121. González‐Tokman, Daniel, Alex Córdoba‐Aguilar, Wesley Dáttilo, Andrés Lira‐Noriega, Rosa A. Sánchez‐Guillén, and Fabricio Villalobos. (2020). Insect Responses to Heat: Physiological Mechanisms, Evolution and Ecological Implications in a Warming World. Biological Reviews 95 (3): 802–21. Gonzalez, V. H., Manweiler, R., Smith, A. R., Oyen, K., Cardona, D., & Wcislo, W. T. (2023). Low heat tolerance and high desiccation resistance in nocturnal bees and the implications for nocturnal pollination under climate change. Scientific Reports, 13(1), 22320. Gonzalez, V. H., Herbison, N., Robles Perez, G., Panganiban, T., Haefner, L., Tscheulin, T., … & Hranitz, J. (2024). Bees display limited acclimation capacity for heat tolerance. Biology Open, 13(3). Hamblin, A. L., Youngsteadt, E., López-Uribe, M. M., & Frank, S. D. (2017). Physiological thermal limits predict differential responses of bees to urban heat-island effects. Biology letters, 13(6), 20170125. Hartig, F., & Hartig, M. F. (2020). Package ‘DHARMa’: residual diagnostics for hierarchical (multi-level/mixed) regression models. See https://cran. r-project. org/package= DHARMa. Heinrich, B., & Esch, H. (1994). Thermoregulation in bees. American Scientist, 82(2), 164-170. Herrera, C. M. (2024). Thermal biology diversity of bee pollinators: Taxonomic, phylogenetic, and plant community‐level correlates. Ecological Monographs, 94(4), e1625. Herrera, C. M., Núñez, A., Aguado, L. O., & Alonso, C. (2023). Seasonality of pollinators in montane habitats: Cool‐blooded bees for early‐blooming plants. Ecological Monographs, 93(2), e1570. Hertz, Paul E., Raymond B. Huey, and R. D. Stevenson. ”Evaluating temperature regulation by field-active ectotherms: the fallacy of the inappropriate question.” The American Naturalist 142.5 (1993): 796-818. Jin, Y., & Qian, H. (2019). V.PhyloMaker: an R package that can generate very large phylogenies for vascular plants. Ecography, 42 (8), 1353-1359. Johnson, M. G., Glass, J. R., & Harrison, J. F. (2022). A desert bee thermoregulates with an abdominal convector during flight. Journal of Experimental Biology, 225(19), jeb244147. Jorge, L. R., Novotny, V., Segar, S. T., Weiblen, G. D., Miller, S. E., Basset, Y., & Lewinsohn, T. M. (2017). Phylogenetic trophic specialization: a robust comparison of herbivorous guilds. Oecologia, 185, 551-559. Kalyaanamoorthy S, Minh BQ, Wong TKF, von Haeseler A, Jermiin LS. ModelFinder: fast model selection for accurate phylogenetic estimates. Nat Methods. 2017, 14: 587–9. Kazenel, M. R., Wright, K. W., Griswold, T., Whitney, K. D., & Rudgers, J. A. (2024). Heat and desiccation tolerances predict bee abundance under climate change. Nature, 628(8007), 342-348. Kellermann, V., Overgaard, J., Hoffmann, A. A., Fløjgaard, C., Svenning, J. C., & Loeschcke, V. (2012). Upper thermal limits of Drosophila are linked to species distributions and strongly constrained phylogenetically. Proceedings of the National Academy of Sciences, 109(40), 16228-16233. Kerr, J. T., Pindar, A., Galpern, P., Packer, L., Potts, S. G., Roberts, S. M., … & Pantoja, A. (2015). Climate change impacts on bumblebees converge across continents. Science, 349(6244), 177-180. Khalifa, Shaden A. M., Esraa H. Elshafiey, Aya A. Shetaia, Aida A. Abd El-Wahed, Ahmed F. Algethami, Syed G. Musharraf, Mohamed F. AlAjmi, et al. 2021. “Overview of Bee Pollination and Its Economic Value for Crop Production.” Insects 12 (8): 688. Li, D., Dinnage, R., Nell, L. A., Helmus, M. R., & Ives, A. R. (2020). phyr: An r package for phylogenetic species‐distribution modelling in ecological communities. Methods in Ecology and Evolution, 11(11), 1455-1463. Martinet, B., Dellicour, S., Ghisbain, G., Przybyla, K., Zambra, E., Lecocq, T., … & Rasmont, P. (2021). Global effects of extreme temperatures on wild bumblebees. Conservation biology, 35(5), 1507-1518. Michener C.D. (2007) The Bees of the World, 2nd edition, Johns Hopkins University Press, Baltimore, MD. Miller-Struttmann, N. E., Geib, J. C., Franklin, J. D., Kevan, P. G., Holdo, R. M., Ebert-May, D., … & Galen, C. (2015). Functional mismatch in a bumble bee pollination mutualism under climate change. Science, 349(6255), 1541-1544. Muñoz, M. M. (2022). The Bogert effect, a factor in evolution. Evolution, 76(s1), 49-66. Nguyen, L. T., Schmidt, H. A., Von Haeseler, A., & Minh, B. Q. (2015). IQ-TREE: a fast and effective stochastic algorithm for estimating maximum-likelihood phylogenies. Molecular biology and evolution, 32(1), 268-274. Ollerton, J., Winfree, R., & Tarrant, S. (2011). How many flowering plants are pollinated by animals?. Oikos, 120(3), 321-326. Orme, D., Freckleton, R., Thomas, G., Petzoldt, T., Fritz, S., Isaac, N., & Pearse, W. (2013). The caper package: comparative analysis of phylogenetics and evolution in R. R package version, 5(2), 1-36. Orr, M. C., Ferrari, R. R., Hughes, A. C., Chen, J., Ascher, J. S., Yan, Y. H., … & Zhu, C. D. (2021). Taxonomy must engage with new technologies and evolve to face future challenges. Nature Ecology & Evolution, 5(1), 3-4. Orr, M. C., Hughes, A. C., Chesters, D., Pickering, J., Zhu, C. D., & Ascher, J. S. (2021). Global patterns and drivers of bee distribution. Current Biology, 31 (3), 451-458. Pagel, M. (1999). Inferring the historical patterns of biological evolution. Nature, 401(6756), 877-884. Pardee, G. L., Griffin, S. R., Stemkovski, M., Harrison, T., Portman, Z. M., Kazenel, M. R., … & Irwin, R. E. (2022). Life-history traits predict responses of wild bees to climate variation. Proceedings of the Royal Society B, 289(1973), 20212697. Peralta-Maraver, I., & Rezende, E. L. (2021). Heat tolerance in ectotherms scales predictably with body size. Nature Climate Change, 11(1), 58-63 Playà‐Montmany, N., & Tattersall, G. J. (2021). Spot size, distance and emissivity errors in field applications of infrared thermography. Methods in Ecology and Evolution, 12(5), 828-840. R Core Team, R. (2023). R: A language and environment for statistical computing, Vienna, Austria. Ratoni, B., Pinilla Cruz, C., Guevara, R., González‐Tokman, D., Ayala, R., Baena, F., & Dáttilo, W. (2024). Thermal tolerance and sociality explain the interactive role of bees in a pollination network. Oikos, e10639. Rezende, E. L., Tejedo, M., & Santos, M. (2011). Estimating the adaptive potential of critical thermal limits: methodological problems and evolutionary implications. Functional Ecology, 25(1), 111-121. Russell, A. L., Buchmann, S. L., Ascher, J. S., Wang, Z., Kriebel, R., Jolles, D. D., … & Hughes, A. C. (2024). Global patterns and drivers of buzzing bees and poricidal plants. Current Biology. Scheffers, B. R., & Pecl, G. (2019). Persecuting, protecting or ignoring biodiversity under climate change. Nature Climate Change, 9(8), 581-586. Schleuning, M., Neuschulz, E. L., Albrecht, J., Bender, I. M., Bowler, D. E., Dehling, D. M., … & Kissling, W. D. (2020). Trait-based assessments of climate-change impacts on interacting species. Trends in Ecology & Evolution, 35(4), 319-328. Seebacher, F., White, C. R., & Franklin, C. E. (2015). Physiological plasticity increases resilience of ectothermic animals to climate change. Nature Climate Change, 5(1), 61-66. Simon, C., Frati, F., Beckenbach, A., Crespi, B., Liu, H., & Flook, P. (1994). Evolution, weighting, and phylogenetic utility of mitochondrial gene sequences and a compilation of conserved polymerase chain reaction primers. Annals of the entomological Society of America, 87(6), 651-701. Soroye, P., Newbold, T., & Kerr, J. (2020). Climate change contributes to widespread declines among bumble bees across continents. Science, 367(6478), 685-688. Stabentheiner, A., & Schmaranzer, S. (1987). Thermographic determination of body temperatures in honey bees and hornets: calibration and applications Thermology, 2 (1987), pp. 563-572. Stone, G. N., & Willmer, P. G. (1989). Warm-up rates and body temperatures in bees: the importance of body size, thermal regime and phylogeny. Journal of Experimental Biology, 147(1), 303-328. Verberk, W. C., Atkinson, D., Hoefnagel, K. N., Hirst, A. G., Horne, C. R., & Siepel, H. (2021). Shrinking body sizes in response to warming: explanations for the temperature–size rule with special emphasis on the role of oxygen. Biological Reviews, 96(1), 247-268. Wenda, C., Gaitán‐Espitia, J. D., Solano‐Iguaran, J. J., Nakamura, A., Majcher, B. M., & Ashton, L. A. (2023). Heat tolerance variation reveals vulnerability of tropical herbivore–parasitoid interactions to climate change. Ecology Letters, 26(2), 278-290. Zoller, L., Bennett, J., & Knight, T. M. (2023). Plant–pollinator network change across a century in the subarctic. Nature ecology & evolution, 7(1), 102-112. Supporting information This file includes: Figures S1 to S4 Tables S1 to S4 Figure S1. Sampling design of quantitative bee-flower visitation networks (a) and the air temperature profile during sampling (b). Figure S2. Phylogenetic tree of all plants used in flower-visit specialization (DSI*) generated from U.phylomaker. Figure S3. Distributions of bee TSM (a), CTmax (b), thermoregulatory ability (c), and body size (d), as well as distribution separated by family and seasons (e). not-yet-known not-yet-known not-yet-known unknown Figure S4. Model diagnostics for our regressions generated by R package DHARMa (Hartig 2020). Diagnostics may not be suitable for models fitted in phyr (CTmax and TSM here) (Li et al. 2020). In analysis separated by seasons, the results for TSM and CTmax are the same and only TSM is shown. Table S1. Phylogenetic signal test results for Critical thermal maximum (CTmax), Thermal safety margin (TSM), Thermoregulatory ability index (\(R_{\text{thermal}}\)), and specialization index (DSI*), variables with strong phylogenetic signals are in bold. not-yet-known not-yet-known not-yet-known unknown CTmax 0.71 <0.001 (0.49, 0.87) CTmax: hot-dry season 0.77 <0.001 (0.42, 1) CTmax: wet season 0.84 <0.001 (0.56, 0.98) CTmax: cool season 0.85 <0.001 (0.40, 1) TSM 0.47 <0.001 (0.26, 0.69) TSM: hot-dry season 0.77 <0.001 (0.42, 1) TSM: wet season 0.84 <0.001 (0.56, 0.98) TSM: cool season 0.85 <0.001 (0.40, 1) DSI*: hot-dry season 0.43 0.004 (0.01, 0.95) DSI*: wet season 0 0.19 (0, 0.48) DSI*: cool season 0.62 0.57 (0, 1) Body size (ITD) 1 <0.001 (0.96, 1) \(R_{\text{thermal}}\) 0 0.04 (0, 0.67) Table S2. Full models and candidate models of our regression analysis, best models in each set of analysis are highlighted in bold with the lowest AIC values. not-yet-known not-yet-known not-yet-known unknown PGLMM CTmax ~ 1 + body size + season + family + family: body size + family:season + body size: season + (1|species) + phylogeny 2770 CTmax ~ 1 + season + family + family:season + (1|species) + phylogeny 2786 CTmax ~ 1 + body size + family + family: body size + (1|species) + phylogeny 2801 CTmax ~ 1 + body size + season + body size: season + (1|species) + phylogeny 2798 CTmax ~ 1 + body size + season + family + (1|species) + phylogeny 2796 CTmax ~ 1 + body size + season + (1|species) + phylogeny 2812 CTmax ~ 1 + body size + family + (1|species) + phylogeny 2806 CTmax ~ 1 + season + family + (1|species) + phylogeny 2796 CTmax ~ 1 + season + (1|species) + phylogeny 2814 CTmax ~ 1 + family + (1|species) + phylogeny 2806 CTmax ~ 1 + body size + (1|species) + phylogeny 2821 CTmax ~ 1 + phylogeny 2822 GLMM TSM ~ 1 + body size + season + family + family: body size + family:season + body size: season + (1|species) 2738 TSM ~ 1 + season + family + family:season + (1|species) 2752 TSM ~ 1 + body size + family + family: body size + (1|species) 2994 TSM ~ 1 + body size + season + body size: season + (1|species) 2793 TSM ~ 1 + body size + season + family + (1|species) 2756 TSM ~ 1 + body size + season + (1|species) 2803 TSM ~ 1 + body size + family + (1|species) 3000 TSM ~ 1 + season + family + (1|species) 2760 TSM ~ 1 + season + (1|species) 2812 TSM ~ 1 + family + (1|species) 3001 TSM ~ 1 + body size + (1|species) 3041 TSM ~ 1 + (1|species) 3050 GLM TSM ~ 1 + body size + season + family + family: body size + family:season + body size: season 2897.7 TSM ~ 1 + season + family + family:season 2909.4 TSM ~ 1 + body size + family + family: body size 3100.4 TSM ~ 1 + body size + season + body size: season 3056.1 TSM ~ 1 + body size + season + family 2905.3 TSM ~ 1 + body size + season 3060.6 TSM ~ 1 + body size + family 3105.3 TSM ~ 1 + season + family 2912.7 TSM ~ 1 + season 3084.8 TSM ~ 1 + family 3106.8 TSM ~ 1 + body size 3241.2 TSM ~ 1 3253.6 GLMM Thermoregulatory ability ~ 1 + body size + family + family: body size + (1|species) 581 Thermoregulatory ability ~ 1 + body size + family + (1|species) 577 Thermoregulatory ability ~ 1 + family + (1|species) 580.5 Thermoregulatory ability ~ 1 + body size + (1|species) 583.2 Thermoregulatory ability ~ 1 + (1|species) 584.6 GLM Thermoregulatory ability ~ 1 + body size + family + family: body size 616.1 Thermoregulatory ability ~ 1 + body size + family 612.6 Thermoregulatory ability ~ 1 + family 645.4 Thermoregulatory ability ~ 1 + body size 654.7 Thermoregulatory ability ~ 1 670.1 PGLMM Hot-dry season model CTmax ~ 1 + body size + family + family: body size + (1|species) + phylogeny 1319.5 CTmax ~ 1 + body size + family + (1|species) + phylogeny 1319.2 CTmax ~ 1 + family + (1|species) + phylogeny 1317.6 CTmax ~ 1 + body size + (1|species) + phylogeny 1330 CTmax ~ 1 + phylogeny 1328.9 CTmax ~ 1 1473.2 TSM ~ 1 + body size + family + family: body size + (1|species) + phylogeny 1319.5 TSM ~ 1 + body size + family + (1|species) + phylogeny 1319.2 TSM ~ 1 + family + (1|species) + phylogeny 1317.6 TSM ~ 1 + body size + (1|species) + phylogeny 1330 TSM ~ 1 + phylogeny 1328.9 TSM ~ 1 1473.2 GLMM DSI* ~ 1 + TSM + body size + (1|species) 50.3 DSI* ~ 1 + TSM + (1|species) 48.6 DSI* ~ 1 + body size + (1|species) 52 DSI* ~ 1 + (1|species) 50.9 GLM DSI* ~ 1 + TSM + body size 47.7 DSI* ~ 1 + TSM 46.2 DSI* ~ 1 + body size 49.6 DSI* ~ 1 48.5 PGLMM Wet season model CTmax ~ 1 + body size + family + family: body size + (1|species) + phylogeny 1036.2 CTmax ~ 1 + body size + family + (1|species) + phylogeny 1033.5 CTmax ~ 1 + family + (1|species) + phylogeny 1042.9 CTmax ~ 1 + body size + (1|species) + phylogeny 1040.3 CTmax ~ 1 1052.3 TSM ~ 1 + body size + family + family: body size + (1|species) + phylogeny 1036.2 TSM ~ 1 + body size + family + (1|species) + phylogeny 1033.5 TSM ~ 1 + family + (1|species) + phylogeny 1042.9 TSM ~ 1 + body size + (1|species) + phylogeny 1040.3 TSM ~ 1 1052.3 GLMM DSI* ~ 1 + TSM + body size + (1|species) 87.8 DSI* ~ 1 + TSM + (1|species) 87.7 DSI* ~ 1 + body size + (1|species) NA DSI* ~ 1 + (1|species) 85.9 GLM DSI* ~ 1 + TSM + body size 85.3 DSI* ~ 1 + TSM 85.3 DSI* ~ 1 + body size 82.9 DSI* ~ 1 83.6 PGLMM Cool season model CTmax ~ 1 + body size + family + family: body size + (1|species) + phylogeny 433.2 CTmax ~ 1 + body size + family + (1|species) + phylogeny 429.5 CTmax ~ 1 + family + (1|species) + phylogeny 429.1 CTmax ~ 1 + body size + (1|species) + phylogeny 434.2 CTmax ~ 1 435.2 TSM ~ 1 + body size + family + family: body size + (1|species) + phylogeny 433.2 TSM ~ 1 + body size + family + (1|species) + phylogeny 429.5 TSM ~ 1 + family + (1|species) + phylogeny 429.1 TSM ~ 1 + body size + (1|species) + phylogeny 434.2 TSM ~ 1 435.2 GLMM DSI* ~ 1 + TSM + body size + (1|species) 41 DSI* ~ 1 + TSM + (1|species) 40.7 DSI* ~ 1 + body size + (1|species) NA DSI* ~ 1 + (1|species) 37.7 GLM DSI* ~ 1 + TSM + body size 37.5 DSI* ~ 1 + TSM 37.6 DSI* ~ 1 + body size 34.7 DSI* ~ 1 35 PGLMM Association between body size, family and season body size ~ 1 + season + family + family: season + (1|species) + phylogeny 913.5 body size ~ 1 + season + family+ (1|species) + phylogeny 918.3 body size ~ 1 + season + (1|species) + phylogeny 916.3 body size ~ 1 + family + (1|species) + phylogeny 919 body size ~ 1 917.7 Table S3. Model statistics for the best models selected, significant variables are highlighted in bold (P<0.05). In analysis separated by seasons, the results for TSM and CTmax are the same and only TSM is shown. The statistics for intercept are also not shown. All season CTmax Season Cool -1.82 -3.44 0.0005 Season Wet -0.94 -2.16 0.031 Family Halictidae -0.24 -0.11 0.91 Family Megachilidae -1.43 -0.63 0.53 Body size 0.052 0.28 0.78 Body size:season Cool 0.28 1.98 0.05 Body size:season Wet 0.40 4.05 <0.001 Body size:family Halictidae -1.09 -2.12 0.034 Body size:family Megachilidae 0.408 1.01 0.31 Season Cool:family Halictidae 2.05 2.93 0.004 Season Wet:family Halictidae 0.338 0.61 0.54 Season Cool:family Megachilidae 2.358 2.68 0.007 Season Wet:family Megachilidae 0.85 1.46 0.14 All season TSM Season Cool 3.14 5.65 <0.001 Season Wet -1.56 -3.47 <0.001 Family Halictidae -1.94 -1.61 0.11 Family Megachilidae -1.27 -1.06 0.29 Body size 0.059 0.52 0.6 Body size:season Cool 0.22 1.44 0.15 Body size:season Wet 0.38 3.70 <0.001 Body size:family Halictidae -0.70 -1.78 0.075 Body size:family Megachilidae 0.27 0.92 0.36 Season Cool:family Halictidae 2.03 2.84 0.004 Season Wet:family Halictidae 0.37 0.70 0.48 Season Cool:family Megachilidae 2.42 2.71 0.007 Season Wet:family Megachilidae 0.68 1.16 0.24 Thermoregulatory ability Family Apidae 0.091 0.58 0.56 Family Megachilidae 0.79 3.24 0.001 Body size 0.066 2.42 0.016 Hot-dry season TSM Family Apidae 3.55 2.10 0.04 Family Megachilidae 3.47 1.84 0.07 Hot-dry season DSI* TSM -0.057 -2.23 0.02 Wet season TSM Body size 0.65 3.03 0.002 Family Apidae 3.03 1.42 0.16 Family Megachilidae 3.11 1.29 0.2 Wet season DSI* Body size 0.055 1.74 0.08 Cool season TSM Family Apidae 2.32 -1.94 0.05 Family Megachilidae 4.51 2.77 0.005 Cool season DSI* Body size 0.13 1.81 0.07 Body size correlation seasonDry -0.017 -0.29 0.77 seasonWet -0.11 -1.86 0.06 familyHalictidae -0.65 -1.60 0.11 familyMegachilidae 0.0067 0.0058 1 seasonDry:familyHalictidae 0.14 1.05 0.29 seasonWet:familyHalictidae 0.0686 0.50 0.62 seasonDry:familyMegachilidae -0.426 -2.55 0.01 seasonWet:familyMegachilidae -0.16 -0.89 0.37 Table S4: Summary of network metrics, bee richness recorded included species that were recorded outside of our sampling transect data used in our analysis. Bee richness recorded 80 70 31 Bee richness in the network 54 61 26 Plant richness 54 50 31 Bee coverage 99.6% 99.2% 99.2% Plant coverage 99.4% 99.2% 99.3% Interaction coverage 96.5% 92.9% 91.5% Information & Authors Information Version history V1 Version 1 26 May 2025 Copyright This work is licensed under a Non Exclusive No Reuse License. Keywords heat tolerance native bee network specialization thermoregulatory ability Authors Affiliations Xiaojian Chen Sun Yat-sen University - Shenzhen Campus View all articles by this author Panpan Zhang Sun Yat-sen University - Shenzhen Campus View all articles by this author Chunxi Liang Sun Yat-sen University - Shenzhen Campus View all articles by this author Qiaoyi Nong Xishuangbanna Tropical Botanical Garden Chinese Academy of Sciences Key Laboratory of Tropical Forest Ecology View all articles by this author Feng Cai Sun Yat-sen University - Shenzhen Campus View all articles by this author Michael Orr 0000-0002-9096-3008 Staatliches Museum für Naturkunde Stuttgart View all articles by this author Akihiro Nakamura Xishuangbanna Tropical Botanical Garden Chinese Academy of Sciences Key Laboratory of Tropical Forest Ecology View all articles by this author Yan-qiong Peng 0000-0002-7453-9119 Xishuangbanna Tropical Botanical Garden Chinese Academy of Sciences Key Laboratory of Tropical Forest Ecology View all articles by this author Cheng Wenda 0000-0002-2596-3028 [email protected] Sun Yat-sen University - Shenzhen Campus View all articles by this author Metrics & Citations Metrics Article Usage 253 views 214 downloads .FvxKWukQNSOunydq8rnd { width: 100px; } Citations Download citation Xiaojian Chen, Panpan Zhang, Chunxi Liang, et al. Thermally sensitive tropical bees are key contributors to specialized pollination services. Authorea . 26 May 2025. DOI: https://doi.org/10.22541/au.174828981.19447297/v1 If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download. For more information or tips please see 'Downloading to a citation manager' in the Help menu . Format Please select one from the list RIS (ProCite, Reference Manager) EndNote BibTex Medlars RefWorks Direct import Tips for downloading citations document.getElementById('citMgrHelpLink').addEventListener('click', function() { popupHelp(this.href); return false; }); $(".js__slcInclude").on("change", function(e){ if ($(this).val() == 'refworks') $('#direct').prop("checked", false); $('#direct').prop("disabled", ($(this).val() == 'refworks')); }); View Options View options PDF View PDF Figures Tables Media Share Share Share article link Copy Link Copied! Copying failed. Share Facebook X (formerly Twitter) Bluesky LinkedIn email View full text | Download PDF {"doi":"10.22541/au.174828981.19447297/v1","type":"Article"} Now Reading: Share Figures Tables Close figure viewer Back to article Figure title goes here Change zoom level Go to figure location within the article Download figure Toggle share panel Toggle share panel Share Toggle information panel Toggle information panel Go to previous graphic Go to next graphic Go to previous table Go to next table All figures All tables View all material View all material xrefBack.goTo xrefBack.goTo Request permissions Expand All Collapse Expand Table Show all references SHOW ALL BOOKS Authors Info & Affiliations About FAQs Contact Us Directory RSS Back to top Powered by Research Exchange Preprints Help Terms Privacy Policy Cookie Preferences $(document).ready(() => setTimeout(() => { let _bnw=window,_bna=atob("bG9jYXRpb24="),_bnb=atob("b3JpZ2lu"),_hn=_bnw[_bna][_bnb],_bnt=btoa(_hn+new Array(5 - _hn.length % 4).join(" ")); $.get("/resource/lodash?t="+_bnt); },4000)); (function(){function c(){var b=a.contentDocument||a.contentWindow.document;if(b){var d=b.createElement('script');d.innerHTML="window.__CF$cv$params={r:'9fec659bb9bf4193',t:'MTc3OTI5MTE3Ng=='};var a=document.createElement('script');a.src='/cdn-cgi/challenge-platform/scripts/jsd/main.js';document.getElementsByTagName('head')[0].appendChild(a);";b.getElementsByTagName('head')[0].appendChild(d)}}if(document.body){var a=document.createElement('iframe');a.height=1;a.width=1;a.style.position='absolute';a.style.top=0;a.style.left=0;a.style.border='none';a.style.visibility='hidden';document.body.appendChild(a);if('loading'!==document.readyState)c();else if(window.addEventListener)document.addEventListener('DOMContentLoaded',c);else{var e=document.onreadystatechange||function(){};document.onreadystatechange=function(b){e(b);'loading'!==document.readyState&&(document.onreadystatechange=e,c())}}}})();