Understanding foraging behavior, and the decisions made by animals when collecting food requires consideration of the energy spent foraging and the energy gained by the nutrient intake (1, 2). However, foraging behavior strategies can be influenced by diverse factors such as life history traits (e.g. parental care, degree of sociality, etc.; 3, 4). Indeed, foraging behavior is even more complex in social animals, where not only the individual fitness has to be considered, but also the fitness of the group (Galef & 5, 6). In eusocial species, such as honeybees and bumblebees, the foraging behavior of the nonreproductive workers is defined primarily by the needs of the colony, rather than by the foraging individual’s physiological needs (7).

Among eusocial bees, colony size can range from some tens of workers for the smallest bumblebee colonies up to several tens of thousands for honeybees (8). In bumblebees, a large colony size is often adaptive. By increasing their colony size, bumblebees can increase their foraging range (9) and thus the amount of resources they can gather (10). Larger colonies can also increase their reproductive success by producing more sexual brood (11, 12) and they are more resilient to stressors (13, 14). Smaller colonies would be more vulnerable to individual errors, as they have fewer additional workers to recruit if individual mistakes during the foraging are made (e.g. a worker failing to efficiently gather resources and return with them to the colony), or if food intake is too low. However, bumblebee colony size is constrained by several factors. Firstly, habitat appears to set strong limits on colony size – for example, some arctic species can have colonies with only 20 workers (15–17), while tropical species can have colonies with more than one thousand workers (15, 16). Secondly, colony size varies temporally, starting with extremely few workers after the colony is first founded by the queen and reaching a peak population by the summer (9). Colony size is also dependent of the availability of resources (9, 10) and parasitism (18). In addition, anthropogenic drivers can lead to selection episodes and colony mortality. Bumblebee colonies are particularly sensitive during the early stages of their life cycle, and stressors such as pesticides (19) and high developmental temperatures (20) can impede colony growth, leading to smaller colonies or to a drastic reduction of an existing colony.

Colonies can adapt to reductions in the number of workers by, for example, extending their growth phase (21). However, the extent to which workers will adapt their foraging behavior to a reduction colony size is mostly unknown. Among the few studies assessing changes in foraging behavior due to colony reduction, Beekman et al. (22) showed that, in honeybees, foraging distance decreased with worker number but that the number of floral patches visited (proportionally to the number of workers in the colony) increased. However, it remains unclear if the proportion of workers recruited to forage was significantly higher in smaller colonies or if, instead, the workers from these smaller colonies were performing significantly more trips per individual. Using the bumblebee Bombus terrestris, Biella et al. (23) highlighted that, despite a reduction of the workforce, the diversity of plants collected remained roughly the same and that the foraging activity of the workers in reduced colonies significantly increased. These few studies could indicate that the foraging behavior of social bees is resilient to drastic reductions in colony size, but the effect of this reduction on a wider range of foraging parameters remains to be investigated.

To better understand how variations in colony size affect foraging behavior, we investigate how reduced colony size influences several foraging parameters in the buff-tailed bumblebee Bombus terrestris. Previous studies have indicated that reductions in worker number can occur as a result of exposure to a stressor (e.g. heat; 20, 24) and that this, in turn, lead to changes in foraging behavior. However, whether these changes were produced by physiological effects of the stressor per se, by colony size reduction, or the combination of these two variables is unknown. We tested the hypothesis that changes in colony size alone (i.e. without any additional stressor) affect bumblebee foraging behavior. To do this, we manipulated the colony size of B. terrestris such that they had a normal number of workers (100) and colonies that had half this number (50). We then placed the colonies in a large flight room and allowed the workers to freely forage on two different types of plants, Borago officinalis and Echium plantagineum. Foraging behavior was analyzed by recording the following parameters: (i) foraging time, (ii) number of foraging trips, (iii) handling time, (iv) visiting rate, (v) number of workers foraging. In agreement with a study performed in field conditions (23), we hypothesize that either the proportion of workers foraging from the reduced colonies, or the number of foraging trips per individual could increase in response to the reduction of the work force. Indeed, Pendrel & Plowright (25) showed that colony size reduction could lead to the increase of the larval feeding rate, and that the behavioral plasticity of bumblebee workers could allow a reallocation of tasks to recruit more foragers (26, 27).

The goal of this study was to assess if a drastic reduction in colony size would affect the foraging behavior of bumblebees. Surprisingly, our analyses showed that, with the exception of the number of foraging trips on E. plantagineum, none of the aspects of foraging behavior we measured were affected by colony size. This is not consistent with previous studies, which have showed that stressors strongly affected several aspects of foraging behavior. For example, exposure to elevated temperature during development caused an increase in worker visiting rate and the number of foraging trips, while the handling time and the foraging time decreased, suggesting that workers were trying to maximize the collection of resources in stressful conditions (20). Interestingly, in this previous work, the number of workers produced in the colonies that were exposed to higher developmental temperatures was reduced, so it was not possible to assess if the changes in foraging behavior were the consequence of a direct effect of temperature on the physiology/behavior of workers, or if the main driver of these changes was the decrease in colony size. The results of this present study suggest that the reduction in colony size alone was not enough to explain changes in foraging behavior observed among bumblebees that developed at an elevated temperature.

When focusing solely on colony size reduction, several studies have indicated that foraging behavior is robust to changes in worker number. Interestingly, in field conditions, Biella et al. (23) highlighted that a removal of 50% of bumblebee workers in a colony lead to an increase in the number of foraging trips of the remaining workers. This increase of activity resulted in the reduced colonies having a similar diet to the non-reduced colonies. Together with our findings, this study suggests that, independently of the experimental conditions (field conditions versus controlled conditions), reduced colonies can compensate for the loss of workers by increasing the foraging activity of their workers. Among honeybees, smaller colonies of Apis mellifera forage on approximately the same number of flower patches as larger colonies by reducing the number of workers visiting each patch (rather than increasing the proportion of foragers; 22). We observed a different pattern in our study: the proportion of workers foraging from the smaller colonies was higher than in the larger colonies. Indeed, while around 20% of the workers from the small colonies were observed visiting the flowers, only 13% of the workers from the larger colonies were visiting them ( Figure 2A ). Thus, while the number of trips per worker in the smaller colonies was higher in only one of the two plant species, the proportion number of workers from smaller colonies recruited to forage was higher overall. The role assigned to workers inside a bumblebee colony is not as fixed as it is in honeybee colonies, and previous studies have shown that bumblebee workers can switch between tasks (25, 36). For example, depending on the resource requirements, some workers can switch from nursing to foraging, although specialized foragers are more efficient at their task and specialized nurses are faster at feeding the brood (26, 37). Even within foragers, Hagbery and Nieh (38) showed that, when some pollen specialist foragers were removed from the colony, generalist foragers could focus on pollen foraging to compensate this lost. We can thus assume that some workers that were not specialized in foraging initially in the smaller colonies became foragers to compensate for the reduced colony size.

Foraging bumblebee workers are also constrained by the quantity of resources brought back in the colony and the quantity of nectar stored in the honey pots – the motivation to forage increases when the quantity of nectar in these pots decreases (39, 40). Interestingly, two studies on honeybees showed that removing pollen stores could increase the proportion of workers foraging in the colony (41, 42). This could explain why we observed a higher proportion of foraging bumblebees in the small colonies in our study: as we did not manipulate the number of cells or larvae, there were still many larvae to feed but a lower worker force to bring back resources, which could have reduced the proportion of pollen stored and lead to the recruitment of more foragers. Indeed, it is known that the size of the brood has an impact on the quantity of pollen brought back to the colony (43). It would thus be interesting to assess if changes in both the size of the brood and the number of workers would ultimately affect the foraging behavior. Another interesting point to assess would be to compare reduced colonies with unmodified colonies. In our framework, we decided to control for the colony size in every colony from the “normal size” treatment, to have a similar number of workers among these colonies (i.e. 100 workers) and logistical limitations prevented us from testing the three treatments simultaneously (i.e. drastically reduced, slightly reduced and unmodified). However, an interesting next step would be to repeat this experiment with half of the colonies unmodified. Based on the results of Biella et al. (23) who did not modify the number of workers in the “normal size” colonies, we could hypothesize that the colonies reduced to 50 workers would compensate for the loss in workforce and have similar foraging parameters to the unmodified colonies.

How colony size impacts the foraging behavior of eusocial insects probably depends on the information used by foragers to localize, reach their food resources and recruit other foragers. For example, Pharaoh’s ants lay pheromone trails to bring back their food sources to the nest. In this species, smaller colonies had a more disorganized foraging system: foragers from these colonies failed to reach a feeder as often as those from larger colonies (44). This impact is probably due to the volatility of the pheromone trails: if too few ants use the trail, it is not maintained and individuals will forage more erratically. In that case, colony size thus had a crucial impact on the foraging behavior and performance. In bumblebees, individual workers must rely primarily on their own ability to locate and learn about profitable floral resources (45). While the efficiency of ant foraging strategy is strongly relying on chemical cues, bumblebee ability to rely primarily on individual learning seem to confer them a higher resilience in their foraging behavior.

As our results indicate that bumblebee foraging behavior is not affected by colony size, the next step would be to assess if the quantity of resources collected and the efficiency of its collection also remain similar. In the desert seed-harvester ants, larger colonies were not more efficient at gathering food from clumped resources, and their foraging rate also remained constant despite variations in colony size (46). This indicates that eusocial insects are quite robust to colony size reduction and are not only able to keep their foraging behavior constant but also their efficiency at collecting resources. However, this assumption remains to be tested in bumblebees, and is particularly important to assess considering that they will be subjected to more stressors caused by human activities in the future. Indeed, while bumblebee foraging behavior was resilient to colony size reduction, other individual or colony features may have been affected by this reduction. For example, Müller and Schmid-Hempel (47), showed that bumblebee colony reduction could affect female body size and decrease the number of males produced. This could ultimately affect the fitness of colonies in the next generation, a hiding cost that should be considered in future experiments.

The original contributions presented in the study are included in the article/ Supplementary Material . Further inquiries can be directed to the corresponding author.

MG and EB conceived the ideas. MG, JM, JZ and EB contributed to the design of the methodology. JM and JZ collected the data. MG, JM and JZ analyzed the data. MG led the writing of the manuscript with critical input from EB. All authors contributed to the article and approved the submitted version.