Tobias J. Smith

Species vulnerability to climate change depends in part on their capacity to evolve in response to increasing heat. Within terrestrial ectotherms, heat tolerance generally corresponds weakly to current climates, which has led many to conclude that this trait is evolutionarily constrained. However, most studies have not accounted for the role of microclimates, potentially obscuring signals of local adaptation. We examine heat tolerance in 95 species of wild bees that vary in nesting behaviour across the latitudinal extent of Australia. We show that species nest (ground, wooden cavities, or plant stems) micro-climate temperatures predict heat tolerance evolution, where stem nesters have evolved the highest heat tolerances, and ground nesters have evolved the lowest heat tolerances due to their ability to evade extreme heat. A moderate level of phylogenetic inertia in heat tolerance is explained by patterns of related species sharing nesting behaviours. This indicates repeated adaptive evolution of similar heat tolerances, rather than strong evolutionary constraints on heat tolerance. Finally, incorporating nesting behaviour into assessments of climate change vulnerability changes the rank order of which species are most at risk. This underscores the need to understand what drives the evolution of heat tolerance across species to better identify the taxa most at risk to climate change.

Nest establishment in stingless bees (Meliponini) is understudied, with observations reported for only about 13 of the approximately 600 species globally. Unlike Apis honey bees (Apini), which start daughter colonies abruptly during a single swarming event, stingless bees establish new nests progressively and daughter colonies receive ongoing support from the parent colony for a period. Here I describe nest establishment for the Australian stingless bee Tetragonula carbonaria in a suburban landscape in South East Queensland, reporting on detailed observations throughout the process at both daughter and parent nests. Daughter nest establishment took 125 days, lasting from early spring to mid-summer. Worker bees travelled between parent and daughter nests, progressively constructing the nest. Throughout, workers carried propolis, wax and resin in their corbicula and honey and pollen internally from parent to daughter nest. A new queen arrived at the daughter nest early in the process, and the daughter colony began to raise young. By the time observable parental support ceased, the daughter colony was a populous, thriving independent colony with extensive propolis structures, food, wax and resin stores, thousands of workers, a reigning queen, had a large brood mass and had completed two full brood rearing cycles. The process reported here is the first for an Australian stingless bee species and while only a single case, it fits the broad understanding of the nest establishment approach taken by stingless bees globally, with a gradual long-term nest construction process with prolonged parental nest support.

Summary: Species vulnerability to climate change depends in part on their capacity to evolve in response to increasing heat ¹ . Within terrestrial ectotherms, heat tolerance generally corresponds weakly to current climates, which has led many to conclude that this trait is evolutionarily constrained ^2–4 . However, most studies have not accounted for the role of microclimates, potentially obscuring signals of local adaptation. We examined heat tolerance in 95 species of wild bees that varied in nesting behaviour across the latitudinal extent of Australia. Species nest (ground, wooden cavities, or plant stems) micro-climate temperatures predicted heat tolerance evolution, where stem nesters evolved the highest heat tolerances, and ground nesters evolved the lowest heat tolerances due to their ability to evade extreme heat. A moderate level of phylogenetic inertia in heat tolerance was explained by patterns of related species sharing nesting behaviours. This indicated repeated adaptive evolution of similar heat tolerances, rather than strong evolutionary constraints on heat tolerance. Finally, incorporating nesting behaviour into assessments of climate change vulnerability changed the rank order of which species were most at risk. This underscores the need to understand what drives the evolution of heat tolerance across species to better identify the taxa most at risk to climate change.

Stingless bees are a widespread group of highly social bees found in tropical regions throughout much of the world. Despite an impressive diversity, relatively little is known about worker behaviour and division of labour. In this study, we investigate the progression of colony tasks over the lifespan of worker bees in colonies of the two most commonly kept Australian species: Tetragonula carbonaria and T. hockingsi. We marked cohorts of 25–100 newly emerged female bees with a paint dot and released them back into colonies housed in observation hives before recording twice weekly behaviours of marked bees within the nest. Foragers were observed through a clear plastic entrance tube. We replicated this with 5–6 marked cohorts across three colonies for each species. We found the two species were similar to each other in the frequency and age distribution of behaviours. Young worker bees were mostly found cleaning, filling or constructing brood cells and collecting honey from food pots. Middle aged bees were more likely to build or maintain food pots or supporting structures, with guarding and foraging occupying the oldest bees. There was, however, significant overlap in timing of tasks. Moreover, fast cohorts progressed to foraging in less than half the time of the slowest cohorts. Despite subtle differences between our Tetragonula species and other stingless bees, it adds to the evidence that progression from safe to risky jobs with age is an ancestral feature shared across stingless bees, and has similarities to honeybees despite an independent evolutionary origin.

Urbanisation drives overall declines in insect pollinators. Although urban green spaces can provide suitable habitat for pollinators much remains to be learned about how urban landscapes either promote or negatively impact pollinators. We investigated how backyard design, local (100 m) and landscape (500 m) scale vegetation cover and human population density were associated with non-eusocial native bee species, eusocial bees (Apis melliferaandTetragonula spp.), and hoverflies, in residential green spaces of the subtropical city Brisbane, Australia. We found that associations between bee abundance and vegetation cover were moderated by human density, but the direction of this effect differed for non-eusocial and eusocial species. Non-eusocial bee abundance was positively associated with tree cover at local and landscape scales when human densities were low, but negatively so at high human population densities. We suggest this may be because the quality of vegetation for non-eusocial bees deteriorates as human density increases. In contrast, abundance of eusocial bees was negatively associated with increasing local cover of grass and shrubs at low levels of human density, but positively associated at high densities. This affinity to humans could partly be explained by domesticated “kept” hives. We found no effect of urban gradients on bee species richness. Hoverfly abundance was negatively related to human density and positively related to vegetation cover at local and landscape scales. At the backyard scale, both bee species richness and bee and hoverfly abundances were positively associated to flower abundance. Backyards with more vegetation cover had higher densities of non-eusocial bees. Our results thus support the idea that urban greening in densely populated areas at multiple spatial scales can benefit a range of insect pollinators.

Two stingless bee species, Tetragonula carbonaria and Tetragonula hockingsi, engage in extreme inter-colony fights, both within and between species. Inter species fights can result in one species taking over the nest of the other. Following successful takeovers, brood from the previous colony could be retained and become workers, but this has yet to be tested. We first tested whether T. carbonaria callows are accepted when introduced into a different conspecific nest and found that they are indeed accepted as normal workers. This suggested that mixed species cohorts might be expected after a hive takeover. We assessed this in a managed T. carbonaria hive, fitted with an observation window that was subject to multiple takeover attempts. Using microsatellite genotyping, we established that multiple different colonies (of both species) attacked the observation colony over 63 days, resulting in a catastrophic collapse of the nest. Our observations inside the hive during these inter-colony invasions revealed no evidence of resource robbing, adding further support to the hypothesis that fighting behaviour in these species has the primary goal of taking over the nest. Newly emerged callows were ejected by invaders, predominantly during the initial few days of fighting. The ejection of callows ceased; however, shortly after each takeover attempt, and, following this, newly emerged callows were accepted as workers. These findings add new evidence that the invading colony co-exists with workers from the previous brood following a takeover, but only until the last of that cohort dies.

(1) Background: Landscape simplification is a major threat to bee and wasp conservation in the tropics, but reliable, long-term population data are lacking. We investigated how community composition, diversity, and abundance of tropical solitary bees and wasps change with landscape simplification (plant diversity, plant richness, distance from forest, forest cover, and land use type) and season. (2) Methods: We installed 336 timber and cob trap nests in four complex forests and three simplified orchards within the subtropical biodiversity hotspot of south-east Queensland, Australia. Trap nests were replaced every season for 23 months and all emergents identified. (3) Results: We identified 28 wasp species and 13 bee species from 2251 brood cells. Bee and wasp community composition changed with landscape simplification such that large, ground-nesting, and spider-hunting species were present in all landscapes, while those with specialist resource requirements and (clepto) parasitoids were present only in complex landscapes. Abundance and diversity of bees and wasps were unaffected by landscape simplification but increased with rainfall. (4) Conclusions: This study highlights the need for multi-year studies incorporating nuanced measures such as composition with a focus on functional diversity to detect changes bee and wasp populations.

Discrepancies in mitochondrial and nuclear genetic data are often interpreted as evidence of hybridisation. We re-examined reports of hybridisation in three cryptic stingless bee species in the genus Tetragonula in South East Queensland, Australia ( T. carbonaria, T. davenporti , and T. hockingsi ). Previous studies on this group using microsatellite markers proposed that occasional hybrids are found. In contrast, we find that allele frequencies at neutral regions of the nuclear genome, both microsatellites and random snps , reliably separated the three species, and thus do not support hybridisation. We found no inter-species variation in PCR amplicons of the nuclear gene EF1alpha , but low and moderate species-specific polymorphisms in the nuclear gene Opsin and the mitochondrial 16S respectively, with no cases of mito-nuclear discordance at these genes. We confirm that nuclear divergence between these species is low, based on 10-26kb of non-coding sequence flanking EF1alpha and Opsin (0.7-1% pairwise difference between species). However, we find mitogenomes to be far more diverged than nuclear genomes (21.6-23.6% pairwise difference between species). Based on these comprehensive analyses of multiple marker types, we conclude that there is no ongoing gene flow in the Tetragonula species of South East Queensland, despite their high morphological similarity to one another and the low nuclear divergence among them. The mitogenomes and draft nuclear genomes provided for these species will be a resource for further molecular studies on this group, which are important pollinators in Australian natural and agroecosystems.

Central to the success of restoration plantings within abandoned pastures is the appropriate selection of species that can establish and grow rapidly to form canopies to suppress grasses. However, species selection can be difficult, largely due to combinations of biotic and abiotic factors operating across multiple spatial scales that can affect seedling growth rates. Using a large replicated restoration experiment in Australia's Wet Tropics, we evaluated seedling growth rates of 24 native rainforest species commonly used in local restoration efforts over the first 31 months post‐planting. We investigated the influence of landscape, site, and planting conditions on early‐stage seedling growth and whether functional traits and surrounding neighborhood density and composition explain variation in seedling growth rates. Seedling growth rates were influenced by numerous stem‐, species‐, plot‐, and climate‐level factors, with the strength of these effects strongly dependent on the size of the seedling. Specifically, species with low wood densities and larger seeds grew faster. In response to plot‐scale and climate factors, larger seedlings consistently displayed faster growth, demonstrating the benefits of initial seedling size for seedling success. Our study highlights that early‐stage seedling growth can be influenced by many factors, operating across multiple spatial scales. Importantly, we demonstrate that planting larger seedlings may improve seedling growth and that developing strategies to increase the survival of fast‐growing low wood density species is crucial for ensuring that plantings can achieve canopy closure quickly, improving early to mid‐term trajectories of tropical forest recovery.