Deep learning for identifying bee species from images of wings and pinned specimens.

One of the most challenging aspects of bee ecology and conservation is species-level identification, which is costly, time consuming, and requires taxonomic expertise. Recent advances in the application of deep learning and computer vision have shown promise for identifying large bumble bee (Bombus) species. However, most bees, such as sweat bees in the genus Lasioglossum, are much smaller and can be difficult, even for trained taxonomists, to identify. For this reason, the great majority of bees are poorly represented in the crowdsourced image datasets often used to train computer vision models. But even larger bees, such as bumble bees from the B. vagans complex, can be difficult to separate morphologically. Using images of specimens from our research collections, we assessed how deep learning classification models perform on these more challenging taxa, qualitatively comparing models trained on images of whole pinned specimens or on images of bee forewings. The pinned specimen and wing image datasets represent 20 and 18 species from 6 and 4 genera, respectively, and were used to train the EfficientNetV2L convolutional neural network. Mean test precision was 94.9% and 98.1% for pinned and wing images respectively. Results show that computer vision holds great promise for classifying smaller, more difficult to identify bees that are poorly represented in crowdsourced datasets. Images from research and museum collections will be valuable for expanding classification models to include additional species, which will be essential for large scale conservation monitoring efforts.

Vertical Stratification of Solitary Bees and Wasps in an Urban Forest from the Brazilian Amazon.

Solitary bees and wasps that nest in cavities in tree trunks are important components of terrestrial ecosystems, providing pollination services, and in the case of wasps, the regulation of their prey populations. However, little is known about the vertical strata where bees and wasps build their nests. This is especially the case of urban forest remnants in the Amazon, which is relevant in the context of the global crisis in insect losses. We investigated the existence of vertical stratification in the nesting of solitary bees and wasps in an urban forest in Rio Branco, state of Acre, in the western Brazilian Amazon. We focused on whether wood temperature, ants, and termites are predictors of bee and wasp nesting. We sampled bee and wasp nests in the forest using trap-nests made with wooden blocks containing cavities with three different diameters for twelve months. Trap-nests were installed randomly at three heights in the forest. We collected 145 nests of 25 species, belonging to 11 genera and 6 families. A higher number of nests and species were collected in the upper stratum of the forest, strengthening the hypothesis that there is vertical stratification in the assemblage of solitary bees and wasps. Wood surface temperature and termite attacks on trap-nests were significantly different between strata, which may explain the vertical stratification of bee and wasp assemblages. Considering the importance of these insects for tropical forest ecosystems, the conservation of structurally complex and stratified forests is of paramount importance to maintain the diversity of this insect group.

Heat and desiccation tolerances predict bee abundance under climate change.

Climate change could pose an urgent threat to pollinators, with critical ecological and economic consequences. However, for most insect pollinator species, we lack the long-term data and mechanistic evidence that are necessary to identify climate-driven declines and predict future trends. Here we document 16 years of abundance patterns for a hyper-diverse bee assemblage1 in a warming and drying region2, link bee declines with experimentally determined heat and desiccation tolerances, and use climate sensitivity models to project bee communities into the future. Aridity strongly predicted bee abundance for 71% of 665 bee populations (species × ecosystem combinations). Bee taxa that best tolerated heat and desiccation increased the most over time. Models forecasted declines for 46% of species and predicted more homogeneous communities dominated by drought-tolerant taxa, even while total bee abundance may remain unchanged. Such community reordering could reduce pollination services, because diverse bee assemblages typically maximize pollination for plant communities3. Larger-bodied bees also dominated under intermediate to high aridity, identifying body size as a valuable trait for understanding how climate-driven shifts in bee communities influence pollination4. We provide evidence that climate change directly threatens bee diversity, indicating that bee conservation efforts should account for the stress of aridity on bee physiology.

Disponibilidad y utilización de los recursos poliníferos por Apis mellifera (Hymenoptera: Apidae) en el este de las Yungas de Jujuy (Argentina) / Availability and use of polliniferous resources by Apis mellifera (Hymenoptera: Apidae) in the Eastern Yungas of Jujuy (Argentina)

Resumen Introducción: El conocimiento de la riqueza vegetal y la estacionalidad alrededor de los apiarios de Apis mellifera es una herramienta de planificación indispensable para los apicultores. Debe incluir la disponibilidad de recursos, las preferencias alimenticias y el comportamiento de búsqueda de alimento. Dicha información no está disponible para las Yungas argentinas, uno de los ecosistemas forestales más estacionales de América del Sur. Objetivo: Evaluar la disponibilidad de recursos tróficos a través de un calendario de floración y su relación con las cargas de polen de A. mellifera en las Yungas. Métodos: En El Fuerte, Jujuy, recolectamos muestras mensuales de septiembre a marzo (2014-2015 y 2015-2016) utilizando trampas de polen. Utilizamos técnicas estandarizadas para los análisis palinológicos e índices de asociación para el uso de recursos. Las fenofases fueron Inicio de floración, Plena floración y Fin de floración. Resultados: Se identificaron 47 especímenes botánicos a nivel de especie y 9 a nivel de género. En ambos períodos hubo una oferta moderada de flores al inicio de la primavera, representada igualmente por plantas arbustivas y herbáceas, con un pico de floración en noviembre. Posteriormente, hubo una caída en la disponibilidad, con un pico de floración nuevamente al final de la temporada. En cinco especies de plantas hubo una asociación de media a alta entre la especie vegetal disponible y la presencia de ésta en el espectro polínico de la muestra de polen corbicular recolectada (Vachellia aroma, Blepharocalyx salicifolius, Cantinoa sp., Vernonanthura sp. y Zanthoxylum coco). Conclusión: En esta región hay una oferta moderada de flores de plantas arbustivas y herbáceas a principios de la primavera, con un pico de floración en noviembre y al final de la temporada. Solo cinco, de casi 50 especies de plantas, muestran una asociación de disponibilidad y uso por parte de las abejas.

Abstract Introduction: Knowledge of vegetation richness and seasonality around Apis mellifera apiaries is an indispensable planning tool for beekeepers. It must include resource availability, food preferences and foraging behaviour. Such information is unavailable for the Argentinian Yungas, one of the most seasonal forest ecosystems in South America. Objective: To assess the availability of trophic resources through a flowering calendar and its relationship with A. mellifera pollen loads in the Yungas. Methods: In El Fuerte, Jujuy, we collected monthly samples from September to March (2014-2015 and 2015-2016) using pollen traps. We used standardized techniques for palynological analyses, and association indices for resource use. The phenophases were Beginning of flowering, Full flowering, and End of flowering. Results: We identified 47 botanical specimens to species level and 9 only to genus. In both periods there was a moderate supply of flowers at the beginning of spring, represented equally by shrub and herbaceous plants, with peak flowering in November. Subsequently, there was a drop in availability, with peak flowering again at the end of the season. In five plant species, there was a medium to high association between the plant species available and their presence in the pollen spectrum of the corbicular pollen samples collected (Vachellia aroma, Blepharocalyx salicifolius, Cantinoa sp., Vernonanthura sp. And Zanthoxylum coco). Conclusion: In this region, there is a moderate supply of shrub and herbaceous plant flowers at the beginning of spring, with peak flowering in November and at the end of the season. Only five, out of nearly 50 plant species, show an association of availability and use by bees.

Sympatric cleptobiotic stingless bees have species-specific cuticular profiles that resemble their hosts.

Stingless bees are the largest group of eusocial pollinators with diverse natural histories, including obligate cleptobionts (genus Lestrimelitta) that completely abandoned flower visitation to rely on other stingless bees for food and nest materials. Species of Lestrimeliita are thought to specialize upon different host species, and deception through chemical similarity has been proposed as a mechanism to explain this phenomenon. In the Yucatan Peninsula of Mexico, Scaptotrigona pectoralis is a species chemically distinct from, and not preferred as a host by, locally widespread Lestrimeliita niitkib; witnessing attacks on S. pectoralis colonies offered the opportunity to test the sensory deception hypothesis to cletoparasitism. Analysis of cuticular profiles revealed that the Lestrimelitta attacking S. pectoralis differed significantly in odour bouquet to L. niitkib and, in contrast, it resembled that of S. pectoralis. Further analyses, including morphometrics, mtDNA barcoding, and the examination of taxonomic features, confirmed the existence of two sympatric Lestrimelitta species. The results give support to the hypothesis of chemical deception as a cleptobiotic strategy in Lestrimelitta sp. This is the first evidence that sympatric cleptobionts of the same genus select hosts in accordance with species-specific cuticular profiles, with possible consequences for ecological adaptation and the evolution of these remarkable organisms and the community of stingless bee hosts.

No statistical evidence that honey bees competitively reduced wild bee abundance in the Munich Botanic Garden-a comment on Renner et al. (2021).

In a recent paper, Renner et al. (Oecologia 195:825-831, 2021) concluded, without supporting statistical evidence, that increased density of managed honey-bee hives between 2019 and 2020 intensified competitive effects of honey bees on non-Apis bee species in the Munich Botanic Garden. Analysis of Renner et al.'s observations revealed that, contrary to their assumption, the change in hive numbers did not statistically alter honey-bee visitation to 29 plant species within or between years. Given this consistency, changes in the proportion of non-Apis bees among visitors of the surveyed plant species between years likely represent their responses to reduced overall availability of floral resources during 2020. Thus, Renner et al.'s observations do not provide convincing evidence that honey bees competitively reduced the abundance of non-Apis bees in the Munich Botanic Garden.

Deep learning increases the availability of organism photographs taken by citizens in citizen science programs.

Citizen science programs using organism photographs have become popular, but there are two problems related to photographs. One problem is the low quality of photographs. It is laborious to identify species in photographs taken outdoors because they are out of focus, partially invisible, or under different lighting conditions. The other is difficulty for non-experts to identify species. Organisms usually have interspecific similarity and intraspecific variation, which hinder species identification by non-experts. Deep learning solves these problems and increases the availability of organism photographs. We trained a deep convolutional neural network, Xception, to identify bee species using various quality of bee photographs that were taken by citizens. These bees belonged to two honey bee species and 10 bumble bee species with interspecific similarity and intraspecific variation. We investigated the accuracy of species identification by biologists and deep learning. The accuracy of species identification by Xception (83.4%) was much higher than that of biologists (53.7%). When we grouped bee photographs by different colors resulting from intraspecific variation in addition to species, the accuracy of species identification by Xception increased to 84.7%. The collaboration with deep learning and experts will increase the reliability of species identification and their use for scientific researches.

Extraction, purification, structural character and biological properties of propolis flavonoids: A review.

Propolis is an aromatic substance which is collected by bees and mixed with bee saliva. The plant sources of propolis are mainly consisted with plant exudates from bark, buds and etc. Flavonoids are secondary metabolites widely found in natural plants, which have a variety of health care functions and are the main active ingredients of propolis. This article summarized the types, active ingredients, pharmacological effects, extraction methods and applications of propolis flavonoids, the aim was to provide the theoretical basis for further research and development of propolis flavonoids.

A new bee species of the genus Dasypoda Latreille (Hymenoptera, Apoidea) from Northwest Africa with comparative remarks on the subgenus Microdasypoda Michez.

North Africa, with its vast array of ecosystems and reliefs, constitutes a remarkable place to explore and describe the diversity of wild bees. In this paper, a new bee species of the genus Dasypoda Latreille (Hymenoptera, Apoidea, Melittidae), D. schwarzi Radchenko et Michez sp. nov., is described from the Atlas Mountains area (Morocco and Tunisia). This species belongs to the subgenus Microdasypoda Michez and is phenotypically related to D. brevicornis Pérez, but differs from all other species of this subgenus by the structure of the male genitals, the metasomal sterna, and by its overall hair colour. A detailed comparative diagnosis of D. schwarzi with the other four species of this subgenus is provided, as well as a key to the males of Microdasypoda, and a correction to the diagnosis of the subgenus. This new species is the fortieth described Dasypoda species and should be looked for in other mountain regions of Northwest Africa, such as in the Algerian Atlas where it could be present.

Taxonomic studies on Dasyproctus Lepeletier & Brullé (Hymenoptera: Crabronidae) from North Vietnam with description of two new species.

Seven species of Dasyproctus Lepeletier & Brullé, 1835 (Hymenoptera: Crabronidae) are recorded from North Vietnam. Of these, D. vietnamensis and D. longi are new species, and D. idrieus (Cameron) and D. pentheri Leclercq are recorded for the first time from the country. Keys to both sexes of Dasyproctus from North Vietnam based on morphological characters are given, and new distributional records are also presented.

Twenty-five new species of mining bees (Hymenoptera: Andrenidae: Andrena) from Israel and the Levant.

Andrena is one of the most diverse bee genera, comprising about 1,600 described species of ground-nesting solitary bees. Many Andrena species are plant specialists, and several taxa have been indicated to be important pollinators of wild and/or crop plants. The Eastern Mediterranean Basin and Israel in particular are one of the main world diversity hotspots of Andrena. Based on extensive examination of museum specimens combined with DNA barcoding, we hereby describe twenty-five Levantine species of Andrena new to science: Andrena anathema Pisanty sp. nov., A. ardentia Pisanty sp. nov., A. asluji Pisanty sp. nov., A. curviocciput Pisanty & Wood sp. nov., A. dividicincta Pisanty sp. nov., A. dorchini Pisanty sp. nov., A. euphorbiae Pisanty sp. nov., A. gageae Wood & Pisanty sp. nov., A. herodesi Pisanty & Wood sp. nov., A. hulae Pisanty sp. nov., A. igraeca Pisanty & Wood sp. nov., A. inusitata Pisanty sp. nov., A. janthinoides Pisanty sp. nov., A. longistilus Pisanty & Wood sp. nov., A. lunaris Pisanty & Wood sp. nov., A. macula Pisanty & Wood sp. nov., A. obtusa Pisanty sp. nov., A. ornithogali Pisanty & Wood sp. nov., A. petrae Wood sp. nov., A. protuber Pisanty sp. nov., A. sulfurea Wood sp. nov., A. turmalina Pisanty & Wood sp. nov., A. veronicae Pisanty & Wood sp. nov., A. veterana Pisanty sp. nov., and A. xera Pisanty sp. nov. We synonymise Andrena edentula Wood with A. tadauchii Gusenleitner syn. nov., and recognise four infraspecific names as valid species: Andrena mediterranea Pisanty & Scheuchl stat. nov., A. mizorhina Warncke stat. nov., A. noacki Alfken sp. resurr. and A. ochraceohirta Alfken sp. resurr. We additionally describe the hitherto unknown sexes of four species, provide new records for fifteen species previously unknown from Israel, and list fourteen taxa whose previously reported presence in Israel is considered erroneous or questionable.

Stingless bee propolis, metformin, and their combination alleviate diabetic cardiomyopathy

Abstract Background and aim: Stingless bee propolis, a resinous compound processed by mandibular secretion of stingless bees, is used for maintenance of hygiene and stability of beehives. Research on stingless bee propolis shows therapeutic properties attributed to polyphenols exhibiting antioxidative, antihyperglycemic and antiischemic effect. However, the cardioprotective effect of stingless bee propolis on diabetic cardiomyopathy is unknown. Methods: Adult male Sprague Dawley rats were randomised to five groups: normal group, diabetic group, diabetic given metformin (DM+M), diabetic given propolis (DM+P) and diabetic given combination therapy (DM+M+P) and treated for four weeks. Body weight, fasting blood glucose, food and water intake were taken weekly. At the end of experiment, biomarkers of oxidative damage were measured in serum and heart tissue. Antioxidants in heart tissue were quantified. Part of left ventricle of heart was processed for histological staining including Haematoxylin and Eosin (H&E) stain for myocyte size and Masson's Trichrome (MT) stain for heart fibrosis and perivascular fibrosis. Results: Propolis alleviated features of diabetic cardiomyopathy such as myocyte hypertrophy, heart fibrosis and perivascular fibrosis associated with improvement in antioxidative status. Conclusion: This study reports beneficial effect of propolis and combination with metformin in alleviating histopathological feature of diabetic cardiomyopathy by modulating antioxidants, making propolis an emerging complementary therapy.

Chromosome Evolution in the Genus Partamona (Apidae: Meliponini), with Comments on B Chromosome Origin.

The genus Partamona includes 33 species of stingless bees, of which 11 were studied cytogenetically. The main goal of this study was to propose a hypothesis about chromosomal evolution in Partamona by combining molecular and cytogenetic data. Cytogenetic analyses were performed on 3 Partamona species. In addition, the molecular phylogeny included mitochondrial sequences of 11 species. Although the diploid number was constant within the genus, 2n = 34, B chromosomes were reported in 7 species. Cytogenetic data showed karyotypic variations related to chromosome morphology and the amount and distribution of heterochromatin and repetitive DNA. The molecular phylogenetic reconstruction corroborated the monophyly of the genus and separated the 2 clades (A and B). This separation was also observed in the cytogenetic data, in which species within each clade shared most of the cytogenetic characteristics. Furthermore, our data suggested that the B chromosome in the genus Partamona likely originated from a common ancestor of the species that have it in clade B and, through interspecific hybridization, it appeared only in Partamona rustica from clade A. Based on the above, Partamona is an interesting genus for further investigations using molecular mapping of B chromosomes as well as for broadening phylogenetic data.

Tropical bee species abundance differs within a narrow elevational gradient.

Insect pollination is among the most essential ecosystem services for humanity. Globally, bees are the most effective pollinators, and tropical bees are also important for maintaining tropical biodiversity. Despite their invaluable pollination service, basic distributional patterns of tropical bees along elevation gradients are globally scarce. Here, we surveyed bees at 100 m elevation intervals from 800 to 1100 m elevation in Costa Rica to test if bee abundance, community composition and crop visitor assemblages differed by elevation. We found that 18 of 24 bee species spanning three tribes that represented the most abundantly collected bee species showed abundance differences by elevation, even within this narrow elevational gradient. Bee assemblages at the two crop species tested, avocado and squash, showed community dissimilarity between high and low elevations, and elevation was a significant factor in explaining bee community composition along the gradient. Stingless bees (Tribe Meliponini) were important visitors to both crop species, but there was a more diverse assemblage of bees visiting avocado compared to squash. Our findings suggest that successful conservation of tropical montane bee communities and pollination services will require knowledge of which elevations support the highest numbers of each species, rather than species full altitudinal ranges.

Why Did the Bee Eat the Chicken? Symbiont Gain, Loss, and Retention in the Vulture Bee Microbiome.

Diet and gut microbiomes are intricately linked on both short and long timescales. Changes in diet can alter the microbiome, while microbes in turn allow hosts to access novel diets. Bees are wasps that switched to a vegetarian lifestyle, and the vast majority of bees feed on pollen and nectar. Some stingless bee species, however, also collect carrion, and a few have fully reverted to a necrophagous lifestyle, relying on carrion for protein and forgoing flower visitation altogether. These "vulture" bees belong to the corbiculate apid clade, which is known for its ancient association with a small group of core microbiome phylotypes. Here, we investigate the vulture bee microbiome, along with closely related facultatively necrophagous and obligately pollinivorous species, to understand how these diets interact with microbiome structure. Via deep sequencing of the 16S rRNA gene and subsequent community analyses, we find that vulture bees have lost some core microbes, retained others, and entered into novel associations with acidophilic microbes found in the environment and on carrion. The abundance of acidophilic bacteria suggests that an acidic gut is important for vulture bee nutrition and health, as has been found in other carrion-feeding animals. Facultatively necrophagous bees have more variable microbiomes than strictly pollinivorous bees, suggesting that bee diet may interact with microbiomes on both short and long timescales. Further study of vulture bees promises to provide rich insights into the role of the microbiome in extreme diet switches. IMPORTANCE When asked where to find bees, people often picture fields of wildflowers. While true for almost all species, there is a group of specialized bees, also known as the vulture bees, that instead can be found slicing chunks of meat from carcasses in tropical rainforests. In this study, researchers compared the microbiomes of closely related bees that live in the same region but vary in their dietary lifestyles: some exclusively consume pollen and nectar, others exclusively depend on carrion for their protein, and some consume all of the above. Researchers found that vulture bees lost some ancestral "core" microbes, retained others, and entered into novel associations with acidophilic microbes, which have similarly been found in other carrion-feeding animals such as vultures, these bees' namesake. This research expands our understanding of how diet interacts with microbiomes on both short and long timescales in one of the world's biodiversity hot spots.

Revision and description of three new species of the Palaearctic subgenus Gulanthidium of the wool carder bee genus Anthidium (Apoidea: Megachilidae: Anthidiini).

The subgenus Anthidium (Gulanthidium) is a species-poor group of similar species, which is distributed across the West Palaearctic, from Morocco in the west to central Asia in the east. Due to some incorrect species determinations in the literature, the taxonomic situation remains to be fully understood. The situation is reviewed with reference to a rich, still unpublished resource of material from throughout the range. Anthidium flavissimum sp. nov. is described from Afghanistan, A. preoccipitale sp. nov. from Iran, and A. occidentale sp. nov. from Morocco, based on a combination of colour, structural traits, genital morphology, and morphometry. The subgenus as treated here now comprises six species for which a key is presented. Some of the species are sympatric. Species of the subgenus Gulanthidium are very close to A. (Proanthidium) and it should be further examined as to whether these two subgenera should be better combined in a single subgenus.

High-resolution (mtDNA) melting analysis for simple and efficient characterization of Africanized honey bee Apis mellifera (Hymenoptera:Apidae).

Analysis of the mtDNA variation in Apis mellifera L. has allowed distinguishing subspecies and evolutionary lineages by means of different molecular methods; from RFLP, to PCR-RFLP and direct sequencing. Likewise, geometric morphometrics (GM) has been used to distinguish Africanized honey bees with a high degree of consistency with studies using molecular information. High-resolution fusion analysis (HRM) allows one to quickly identify sequence polymorphisms by comparing DNA melting curves in short amplicons generated by real-time PCR (qPCR). The objective of this work was to implement the HRM technique in the diagnosis of Africanization of colonies of A. mellifera from Argentina, using GM as a validation method. DNA was extracted from 60 A. mellifera colonies for mitotype identification. Samples were initially analyzed by HRM, through qPCRs of two regions (485 bp/385 bp) of the mitochondrial cytochrome b gene (cytb). This technique was then optimizing to amplify a smaller PCR product (207 bp) for the HRM diagnosis for the Africanization of colonies. Of the 60 colony samples analyzed, 41 were classified as colonies of European origin whereas 19 revealed African origin. All the samples classified by HRM were correctly validated by GM, demonstrating that this technique could be implemented for a rapid identification of African mitotypes in Apis mellifera samples.

Machine learning approach for automatic recognition of tomato-pollinating bees based on their buzzing-sounds.

Bee-mediated pollination greatly increases the size and weight of tomato fruits. Therefore, distinguishing between the local set of bees-those that are efficient pollinators-is essential to improve the economic returns for farmers. To achieve this, it is important to know the identity of the visiting bees. Nevertheless, the traditional taxonomic identification of bees is not an easy task, requiring the participation of experts and the use of specialized equipment. Due to these limitations, the development and implementation of new technologies for the automatic recognition of bees become relevant. Hence, we aim to verify the capacity of Machine Learning (ML) algorithms in recognizing the taxonomic identity of visiting bees to tomato flowers based on the characteristics of their buzzing sounds. We compared the performance of the ML algorithms combined with the Mel Frequency Cepstral Coefficients (MFCC) and with classifications based solely on the fundamental frequency, leading to a direct comparison between the two approaches. In fact, some classifiers powered by the MFCC-especially the SVM-achieved better performance compared to the randomized and sound frequency-based trials. Moreover, the buzzing sounds produced during sonication were more relevant for the taxonomic recognition of bee species than analysis based on flight sounds alone. On the other hand, the ML classifiers performed better in recognizing bees genera based on flight sounds. Despite that, the maximum accuracy obtained here (73.39% by SVM) is still low compared to ML standards. Further studies analyzing larger recording samples, and applying unsupervised learning systems may yield better classification performance. Therefore, ML techniques could be used to automate the taxonomic recognition of flower-visiting bees of the cultivated tomato and other buzz-pollinated crops. This would be an interesting option for farmers and other professionals who have no experience in bee taxonomy but are interested in improving crop yields by increasing pollination.

Phylogenomics reveals accelerated late Cretaceous diversification of bee flies (Diptera: Bombyliidae).

Bombyliidae is a very species-rich and widespread family of parasitoid flies with more than 250 genera classified into 17 extant subfamilies. However, little is known about their evolutionary history or how their present-day diversity was shaped. Transcriptomes of 15 species and anchored hybrid enrichment (AHE) sequence captures of 86 species, representing 94 bee fly species and 14 subfamilies, were used to reconstruct the phylogeny of Bombyliidae. We integrated data from transcriptomes across each of the main lineages in our AHE tree to build a data set with more genes (550 loci versus 216 loci) and higher support levels. Our overall results show strong congruence with the current classification of the family, with 11 out of 14 included subfamilies recovered as monophyletic. Heterotropinae and Mythicomyiinae are successive sister groups to the remainder of the family. We examined the evolution of key morphological characters through our phylogenetic hypotheses and show that neither the "sand chamber subfamilies" nor the "Tomophthalmae" are monophyletic in our phylogenomic analyses. Based on our results, we reinstate two tribes at the subfamily level (Phthiriinae stat. rev. and Ecliminae stat. rev.) and we include the genus Sericosoma Macquart (previously incertae sedis) in the subfamily Oniromyiinae, bringing the total number of bee fly subfamilies to 19. Our dating analyses indicate a Jurassic origin of the family (165-194 Ma), with the sand chamber evolving early in bee fly evolution, in the late Jurassic or mid-Cretaceous (100-165 Ma). We hypothesize that the angiosperm radiation and the hothouse climate established during the late Cretaceous accelerated the diversification of bee flies, by providing an expanded range of resources for the parasitoid larvae and nectarivorous adults.