VespAI: a deep learning-based system for the detection of invasive hornets.

The invasive hornet Vespa velutina nigrithorax is a rapidly proliferating threat to pollinators in Europe and East Asia. To effectively limit its spread, colonies must be detected and destroyed early in the invasion curve, however the current reliance upon visual alerts by the public yields low accuracy. Advances in deep learning offer a potential solution to this, but the application of such technology remains challenging. Here we present VespAI, an automated system for the rapid detection of V. velutina. We leverage a hardware-assisted AI approach, combining a standardised monitoring station with deep YOLOv5s architecture and a ResNet backbone, trained on a bespoke end-to-end pipeline. This enables the system to detect hornets in real-time-achieving a mean precision-recall score of ≥0.99-and send associated image alerts via a compact remote processor. We demonstrate the successful operation of a prototype system in the field, and confirm its suitability for large-scale deployment in future use cases. As such, VespAI has the potential to transform the way that invasive hornets are managed, providing a robust early warning system to prevent ingressions into new regions.

Quantifying the impact of an invasive Hornet on Bombus terrestris Colonies.

The invasive hornet Vespa velutina nigrithorax is considered a proliferating threat to pollinators in Europe and Asia. While the impact of this species on managed honey bees is well-documented, effects upon other pollinator populations remain poorly understood. Nonetheless, dietary analyses indicate that the hornets consume a diversity of prey, fuelling concerns for at-risk taxa. Here, we quantify the impact of V. velutina upon standardised commercially-reared colonies of the European bumblebee, Bombus terrestris terrestris. Using a landscape-scale experimental design, we deploy colonies across a gradient of local V. velutina densities, utilising automated tracking to non-invasively observe bee and hornet behaviour, and quantify subsequent effects upon colony outcomes. Our results demonstrate that hornets frequently hunt at B. terrestris colonies, being preferentially attracted to those with high foraging traffic, and engaging in repeated-yet entirely unsuccessful-predation attempts at nest entrances. Notably however, we show that B. terrestris colony weights are negatively associated with local V. velutina densities, indicating potential indirect effects upon colony growth. Taken together, these findings provide the first empirical insight into impacts on bumblebees at the colony level, and inform future mitigation efforts for wild and managed pollinators.

A derived honey bee stock confers resistance to Varroa destructor and associated viral transmission.

The ectoparasite Varroa destructor is the greatest threat to managed honey bee (Apis mellifera) colonies globally. Despite significant efforts, novel treatments to control the mite and its vectored pathogens have shown limited efficacy, as the host remains naïve. A prospective solution lies in the development of Varroa-resistant honey bee stocks, but a paucity of rigorous selection data restricts widespread adoption. Here, we characterise the parasite and viral dynamics of a Varroa-resistant honey bee stock, designated 'Pol-line', using a large-scale longitudinal study. Results demonstrate markedly reduced Varroa levels in this stock, diminished titres of three major viruses (DWV-A, DWV-B, and CBPV), and a two-fold increase in survival. Levels of a fourth virus that is not associated with Varroa-BQCV-do not differ between stocks, supporting a disruption of the transmission pathway. Further, we show that when decoupled from the influence of Varroa levels, viral titres do not constitute strong independent predictors of colony mortality risk. These findings highlight the need for a reassessment of Varroa etiology, and suggest that derived stocks represent a tractable solution to the Varroa pandemic.

An Efficient Alternative to the CDC Gravid Trap for Southern House Mosquito (Diptera: Culicidae) Surveillance.

West Nile virus (WNV) is the most prevalent arbovirus found throughout the United States. Surveillance of surface breeding Culex vectors involved in WNV transmission is primarily conducted using CDC Gravid traps. However, anecdotal claims from mosquito abatement districts in Louisiana assert that other trap types may be more suited to WNV surveillance. To test the validity of these assertions, we conducted a series of trapping trials and WNV surveillance over 3 yr to compare the efficacy of multiple trap types. First, we compared the CDC Gravid trap, CO2-baited New Standard Miniature Blacklight traps, and CO2-baited CDC light traps with either an incandescent light, a red light, or no light. We found that the CDC Gravid trap and CO2-baited no-light CDC Light trap collected the most mosquitoes. Second, we conducted additional, long-term trapping and WNV surveillance to compare these two trap types. We found that CO2-baited no-light CDC traps collected more of the local WNV vector, Culex quinquefasciatus (Say, Diptera, Culicidae), and detected WNV with greater sensitivity. Finally, we conducted trapping to compare the physiological states of Cx. quinquefasciatus and diversity of collected mosquitoes. CO2-baited no-light CDC light traps collected more unfed Cx. quinquefasciatus while Gravid traps collected more blooded Cx. quinquefasciatus; both traps collected the same number of gravid Cx. quinquefasciatus. Additionally, we found that CO2-baited no-light CDC light traps collected a larger diversity of mosquito species than Gravid traps.

Honeybees show a context-dependent rightward bias.

Lateralized behaviour in social insects is of biological significance, as certain lateral biases appear to have emerged in tandem with eusociality, and thus can provide insights into its functioning. Here, I investigate behavioural asymmetry in an ecologically important social insect, the honeybee Apis mellifera. Experiments show that foraging bees exhibit a strong rightward turning bias, accompanied by reduced decision latency when entering open cavities, yet demonstrate no directional preference in sequential choice-mazes. A rightward exploration preference within unknown cavities is consistent with current information relating to the physiology of this species, with workers being better equipped for sensory investigation and threat response using their right antenna and eye. Furthermore, when applied to collective nest-choice scenarios, a similar bias would promote the uniform assessment of nest cavities, and consistency in quorum attainment. Conversely, such laterality appears to provide no immediate advantage in enclosed decision-maze systems, where thigmotaxis instead predominates. As such, my results show that directional biases in A. mellifera are extent, yet context-dependent, thus providing a simple and optimized response to varied social challenges.

Variability in individual assessment behaviour and its implications for collective decision-making.

Self-organized systems of collective behaviour have been demonstrated in a number of group-living organisms. There is, however, less research relating to how variation in individual assessments may facilitate group decision-making. Here, we investigate this using the decentralized system of collective nest choice behaviour employed by the ant Temnothorax albipennis, combining experimental results with computational modelling. In experiments, isolated workers of this species were allowed to investigate new nest sites of differing quality, and it was found that for any given nest quality, there was wide variation among individuals in the durations that they spent within each nest site. Additionally, individual workers were consistent in spending more time in nest sites of higher quality, and less time in those of lower quality. Hence, the time spent in a new nest site must have included an assessment of nest quality. As nest site visit durations (henceforth termed assessment durations) are linked to recruitment, it is possible that the variability we observed may influence the collective decision-making process of colonies. Thus, we explored this further using a computational model of nest site selection, and found that heterogeneous nest assessments conferred a number of potential benefits. Furthermore, our experiments showed that nest quality assessments were flexible, being influenced by experience of prior options. Our findings help to elucidate the potential mechanisms underlying group behaviour, and highlight the importance of heterogeneity among individuals, rather than precise calibration, in shaping collective decision-making.

A social mechanism facilitates ant colony emigrations over different distances.

Behavioural responses enable animals to react rapidly to fluctuating environments. In eusocial organisms, such changes are often enacted at the group level, but may be organised in a decentralised fashion by the actions of individuals. However, the contributions of different group members are rarely homogeneous, and there is evidence to suggest that certain 'keystone' individuals are important in shaping collective responses. Accordingly, investigations of the dynamics and structuring of behavioural changes at both the group and individual level are crucial for evaluating the relative influence of different individuals. Here, we examined the composition of tandem running behaviour during colony emigrations in the ant species Temnothorax albipennis Tandem running is modulated in response to emigration distance, with more runs being conducted when a more distant nest site must be reached. We show that certain individuals are highly active in the tandem running process, attempting significantly more work in the task. Contrary to expectations, however, such individuals are in fact no more successful at conducting tandem runs than their less active nest mates. Instead, it seems that when more tandem runs are required, colonies rely on greater recruitment of workers into the process. The implications of our study are that in some cases, even when apparently 'key' individuals exist within a group, their relative contribution to task performance may be far from decisive.

Migration control: a distance compensation strategy in ants.

Migratory behaviour forms an intrinsic part of the life histories of many organisms but is often a high-risk process. Consequently, varied strategies have evolved to negate such risks, but empirical data relating to their functioning are limited. In this study, we use the model system of the house-hunting ant Temnothorax albipennis to demonstrate a key strategy that can shorten migration exposure times in a group of social insects. Colonies of these ants frequently migrate to new nest sites, and due to the nature of their habitat, the distances over which they do so are variable, leading to fluctuating potential costs dependent on migration parameters. We show that colonies of this species facultatively alter the dynamics of a migration and so compensate for the distance over which a given migration occurs. Specifically, they achieve this by modulating the rate of 'tandem running', in which workers teach each other the route to a new nest site. Using this method, colonies are able to engage a larger number of individuals in the migration process when the distance to be traversed is greater, and furthermore, the system appears to be based on perceived encounter rate at the individual level. This form of decentralised control highlights the adaptive nature of a behaviour of ecological importance, and indicates that the key to its robustness lies in the use of simple rules. Additionally, our results suggest that such coordinated group reactions are central to achieving the high levels of ecological success seen in many eusocial organisms.

Differentiated Anti-Predation Responses in a Superorganism.

Insect societies are complex systems, displaying emergent properties much greater than the sum of their individual parts. As such, the concept of these societies as single 'superorganisms' is widely applied to describe their organisation and biology. Here, we test the applicability of this concept to the response of social insect colonies to predation during a vulnerable period of their life history. We used the model system of house-hunting behaviour in the ant Temnothorax albipennis. We show that removing individuals from directly within the nest causes an evacuation response, while removing ants at the periphery of scouting activity causes the colony to withdraw back into the nest. This suggests that colonies react differentially, but in a coordinated fashion, to these differing types of predation. Our findings lend support to the superorganism concept, as the whole society reacts much like a single organism would in response to attacks on different parts of its body. The implication of this is that a collective reaction to the location of worker loss within insect colonies is key to avoiding further harm, much in the same way that the nervous systems of individuals facilitate the avoidance of localised damage.

Computational model of collective nest selection by ants with heterogeneous acceptance thresholds.

Collective decision-making is a characteristic of societies ranging from ants to humans. The ant Temnothorax albipennis is known to use quorum sensing to collectively decide on a new home; emigration to a new nest site occurs when the number of ants favouring the new site becomes quorate. There are several possible mechanisms by which ant colonies can select the best nest site among alternatives based on a quorum mechanism. In this study, we use computational models to examine the implications of heterogeneous acceptance thresholds across individual ants in collective nest choice behaviour. We take a minimalist approach to develop a differential equation model and a corresponding non-spatial agent-based model. We show, consistent with existing empirical evidence, that heterogeneity in acceptance thresholds is a viable mechanism for efficient nest choice behaviour. In particular, we show that the proposed models show speed-accuracy trade-offs and speed-cohesion trade-offs when we vary the number of scouts or the quorum threshold.

Effects of the neonicotinoid pesticide thiamethoxam at field-realistic levels on microcolonies of Bombus terrestris worker bumble bees.

Neonicotinoid pesticides are currently implicated in the decline of wild bee populations. Bumble bees, Bombus spp., are important wild pollinators that are detrimentally affected by ingestion of neonicotinoid residues. To date, imidacloprid has been the major focus of study into the effects of neonicotinoids on bumble bee health, but wild populations are increasingly exposed to alternative neonicotinoids such as thiamethoxam. To investigate whether environmentally realistic levels of thiamethoxam affect bumble bee performance over a realistic exposure period, we exposed queenless microcolonies of Bombus terrestris L. workers to a wide range of dosages up to 98 µgkg(-1) in dietary syrup for 17 days. Results showed that bumble bee workers survived fewer days when presented with syrup dosed at 98 µg thiamethoxamkg(-1), while production of brood (eggs and larvae) and consumption of syrup and pollen in microcolonies were significantly reduced by thiamethoxam only at the two highest concentrations (39, 98 µgkg(-1)). In contrast, we found no detectable effect of thiamethoxam at levels typically found in the nectars of treated crops (between 1 and 11 µgkg(-1)). By comparison with published data, we demonstrate that during an exposure to field-realistic concentrations lasting approximately two weeks, brood production in worker bumble bees is more sensitive to imidacloprid than thiamethoxam. We speculate that differential sensitivity arises because imidacloprid produces a stronger repression of feeding in bumble bees than thiamethoxam, which imposes a greater nutrient limitation on production of brood.