Bumble bee movement ecology: foraging and dispersal across castes and life stages

Abstract

Movement is a dynamic process that changes with ontogeny, physiological state, and ecological context. The results of organismal movement impact multiple dimensions of fitness, population dynamics, and functional interactions. As such, the study of movement is critical for understanding and conserving species. Bumble bees (Apidae: Bombus spp.) offer a powerful system to study multiple complexities of movement within a functionally important clade. Their life history includes distinct social and solitary phases, substantial intraspecific variation in body size, and multiple modes of movement behavior. These traits allow investigations of diverse concepts at multiple scales and during contrasting behavioral and motivational states—from individuals, to colonies, to populations, and among species. Despite extensive study as model organisms of fine-scale movements and optimal foraging theory, understanding of landscape-scale movements is more limited. This knowledge gap is especially troubling given global pollinator declines because such dispersive movements fundamentally affect how populations respond to landscape transformation, climate change, and restoration efforts. To build toward a refined understanding of the bumble bee movement, inform research, and assist conservation programs, we review foraging and dispersal movement across life stages and castes. Using an ontogenetic approach, we compare the movement motivation and capacity of individuals throughout colony development. Despite the growth in recent literature, much remains to be learned about the bumble bee movement, especially dispersive life stages. Focused effort on how movement varies with individual state such as nutrition and age, and comparative studies of species would all fill knowledge gaps with high potential to improve bee conservation and research.

Introduction

The way organisms move through and interact with their environment profoundly impacts resource acquisition, mating, and species interactions, with implications for population dynamics, community composition, and the functioning of ecosystems (Charnov 1976, Hamilton and May 1977, Hanski 1998, Ricketts 2001, Kremen et al. 2007, Nathan et al. 2008). Summarizing general principles for animal movement is complicated because movement can vary dramatically among closely related species and individuals and movement propensity, range, and behavior often change with ontogeny and development (Nathan et al. 2008, Osborne et al. 2013). These changes may be components of a stereotypical developmental program, such as the transition from a relatively immobile to a highly vagile life-history stage, or they may also be plastic responses to environmental cues, such as the drastic phase-changes of locust populations (Simpson et al. 1999) or differences in the distribution of the resources in a landscape (Woodgate et al. 2016). Movement also can vary among individuals within seemingly homogeneous populations (including a single developmental stage) because of differences in size, internal physiological state, and motivation (Holyoak et al. 2008) or experience and learning (Osborne et al. 2013, Woodgate et al. 2017). For many organisms, we lack a basic understanding of the motivations driving individuals to move, and how resulting movements scale up to populations and metapopulations (Holyoak et al. 2008). Despite these limitations, we nonetheless must choose scales over which to implement conservation programs or conduct ecological studies. Thus, understanding the fundamental principles of an organism’s movement allows us to select more appropriate scales for addressing questions in ecology, evolution, and behavior and to effectively implement management plans.

Bumble bees (Apidae: Bombus spp.) offer a particularly useful system to examine general questions about movement ecology with relevance to many organisms. Their complex life cycle, involving solitary and social stages with the division of labor between reproductive and non-reproducing individuals (Fig. 1), combined with a central-place foraging habit, means that different types of movement (eg foraging, mate location, nest searching, and dispersal) contribute to individual, population, and community processes at different scales. The caste and stage-structured life history of bumble bees also allow these different types of movement to be studied and interpreted relatively independently from each other. For example, although queens and males forage, pollen, and nectar collection during peak colony growth is carried out by non-reproducing workers whose primary motivation is to collect resources for the colony’s brood. This compartmentalization of foraging movement from mate location, nest location, and dispersal has allowed for a simpler interpretation of its motivations, behavioral elements, and fitness consequences (Heinrich 1979). Furthermore, although several of the aspects mentioned above could equally apply to other social hymenoptera, such as the well-studied honey bee (Apis mellifera), bumble bees have several clear distinctions. Notably, bumble bees offer a system to investigate movement across multiple scales of organization from individuals to colonies. For instance, foraging behavior and body size often vary significantly within a colony (Fitzgerald et al. 2022), allowing for investigations of interactions among individual phenotype, ontogeny, and environmental factors and their resultant influence on colony-level patterns of space use.

Fig. 1.

Typical bumble bee life cycle with movement phases. Foundresses emerge from diapause to establish new colonies (1). After locating a nest, queens forage to provision their first brood (2), after which they assume a mainly reproductive role while their daughter-workers carry out the tasks of foraging and brood care (3). Workers are produced in successive cohorts until a demographic switch point after which the reproductive castes, gynes, and males, are produced (4). These individuals leave their natal colonies to mate with individuals from other colonies (4→5). Mated queens then disperse to locate hibernacula and begin overwintering until the following spring (6). Potentially dispersive phases (blue, squiggly, dashed lines) occur three times for queens (1→2, 4→5, 5→6) and for males as they search for mates (4→5). Central-place foraging (orange, circular, dashed lines) occurs during nest building (2), and through subsequent worker cohorts (3). Illustrations by Nicki Bailey.

Bumble bee movement is also particularly relevant for pollinator conservation programs.

Because bumble bees are some of the best documented and easily identifiable bee taxa, their use as indicator species potentially allows for the interpretation of patterns in overall bee declines. Reported declines of bumble bee populations in Europe and North America are alarming (Goulson et al. 2008, Grixti et al. 2009, Cameron et al. 2011, Cameron and Sadd 2020) and many of the main drivers are shared with those of other bee clades. Bumble bee species show clear sensitivity to landscape changes (Bommarco et al. 2012) and habitat fragmentation (Redhead et al. 2016, Gómez-Martínez et al. 2020, Clake et al. 2022), the loss of suitable forage (Carvell et al. 2006), climate-induced range contractions (Kerr et al. 2015, Jackson et al. 2022), agrochemicals (Rundlöf et al. 2015), and introduced pests and parasites (Plischuk et al. 2009, Li et al. 2012, Figueroa et al. 2023). Furthermore, because bumble bees are highly mobile generalist foragers that access resources from diverse habitats, they could be a useful proxy for assessing human impacts on bees generally (Kremen et al. 2002, 2007, Williams et al. 2010). Despite a rapidly growing body of literature on bumble bee ecology and conservation (Cameron and Sadd 2020), our knowledge of their movement lags far behind the need for information in wild and managed systems that rely upon their pollination services. By providing a framework for how to interpret bumble bee movement propensity, behavior, and scale across life cycles, we can ensure that conservation plans are implemented with relevant scales and timing that match expected patterns of foraging and dispersal.

In this review, we propose adaptive explanations for the timing, motivation, scale, and behavioral differences among individuals or ontogenetic phases to understand the bumble bee movement. This approach allows us to examine how movement and scale change throughout the life cycle, and suggests how future studies and management efforts may be informed by considering the varied requirements of different ontogenetic phases. We explicitly focus on movements that have consequences for resource acquisition, population dynamics, and/or the genetic composition of colonies and populations. As such, we give only brief attention to studies focused on the underlying mechanisms of bee movement, such as longevity, navigational techniques, and flight biomechanics, focusing instead on the resultant processes that dictate rates of colonization, genetic drift, and movement propensity. Furthermore, flower visitation at very short temporal scales and/or small spatial scales (ie movement within patches, inflorescences, etc.) has received extensive study (eg Pyke 1978, Heinrich 1979, Schmid-Hempel 1984, 1985, Williams and Thomson 1998, Schulke and Waser 2001) and warrants an independent assessment on its own. Thus, we do not discuss such movements at length here.

Bumble Bee Movement by Life Stage

Free-living bumble bee species can be described as annual species with distinct solitary and social phases (Fig. 1; though note exceptions for socially parasitic species or tropical lineages). Colonies are founded by solitary queens in the spring. We refer to these queens as foundresses. Foundresses may disperse to find nesting sites and begin foraging for a first cohort of daughters as solitary individuals. After these daughters are eclose as adults, they assume in-nest and foraging roles, soon taking over all foraging from their mother. Colonies grow through successive worker cohorts and switch to the production of males and new queens. We refer to these new queens as gynes to aid in distinction from spring foundresses. Emerging males and gynes leave the natal colony to find mates. Mated gynes locate overwintering sites where they remain in diapause until the following spring. Only these new gynes overwinter to become foundresses the following season.

Below, we discuss movement across the life stages and castes broken down into discussions of timing and motivation, propensity, potential and behaviors, spatial scale, and landscape interactions. We make a distinction between potential movement, the distance an individual can hypothetically move under optimal conditions, and realized movement, the actual distance an individual moves after all environmental constraints have been considered (Fig. 2; Kendall et al. 2022).

Fig. 2.

Conceptual diagram of potential (light shading) and realized (darker shading) movements of bumble bee life stages and castes. Dispersive movements (blue, or darker shading when viewed in grayscale) are generally larger than foraging movements (yellow, or lighter shading when viewed in grayscale), and although movement potential should typically remain unchanged with ontogeny (left to right), realized movement should change given various life stage requirements and shifts between dispersive and foraging modes. Movement potential is hypothesized to be driven by an association with body size and movement mode (ie dispersal or central-place foraging), with dispersive phases allowing for greater movement as individuals do not need to return to the nest. Queens should have the greatest movement potential. Whereas males and workers may have similar movement potential based on size alone, the lack of central-place foraging in males allows for greater net displacement.

Queens—Gynes and Foundresses

Introduction

Colonies typically produce queens late in the flight season. Queen’s large nutritional reserves (Alford 1969, Röseler and Röseler 1986) and large body sizes (Greenleaf et al. 2007, Kendall et al. 2022), mean they may be capable of much greater movements than are other stages like workers or males. However, when we consider their motivations to move their realized distances may not always be larger (Fig. 2). In addition to being larger, queens are more behaviorally complex than workers and males. Therefore, changes in movement motivation and pattern likely arise as they mature through stages of mating, overwintering, nest founding, foraging, and reproduction.

Queens are categorized into two ontogenetic phases: gynes, new queens from the time they eclose until they emerge from diapause, and foundresses, mated queens who emerge from diapause, locate nest sites, and found a new colony (Fig. 1). Gynes and foundresses are the same individuals, thus their potential for movement in terms of biomechanical ability, as well as previous environmental impacts are shared. However, their behavior and realized scale of movement may differ dramatically due to differences in internal motivation, physiology, nutritional state, and ecological context.

Gynes are a cryptic phase of the bumble bee life cycle and are rarely observed (relative to workers) in the wild, despite evidence that single colonies can produce dozens to hundreds of them (Macfarlane et al. 1994, Williams et al. 2011). Although the paucity of direct observations makes verified claims of dispersal difficult, the prevailing opinion is that gynes leave the natal nest within a few days after emerging, and do not typically return to the colony. They feed outside the nest to build lipid stores (Alford 1969), mate, and then shortly after enter diapause (Sladen 1912, Plath 1934). Alternatively, there are a few observations of gynes returning to their natal nest with pollen loads (Sladen 1912, Plath 1934, Allen et al. 1978) when sibling-gynes are still developing, and the mother-queen has died or the colony is otherwise resource stressed (Allen et al. 1978). Allen et al. (1978) also observed gynes repeatedly return to the nest without any resources, but before the queen had died. It has been speculated these gynes may have been using the nest as an overnighting location while continuing their searches for mating and overwintering sites (Allen et al. 1978, Goulson 2003). Clearly, this phase of the life cycle deserves more careful study. Given that extended flight for dispersal requires energy that might otherwise be used to maintain queens over the winter, and exposes gynes to predation risk, we posit that long-distance dispersal of gynes should only be likely when suitable mates or overwintering substrate are locally unavailable. The gyne’s choice of an overwinter site is likely to affect overwinter survival strongly. Thus, movement associated with the selection of these sites probably plays a critical role in population dynamics.

Foundresses emerge in the spring already mated and having survived overwintering (though a small number may overwinter without mating, eg Mullins et al. 2020). Given attrition during winter, foundress numbers are fewer than gynes. As such, their movements have a disproportionate influence on bumble bee population structure compared to other life stages (Berg et al. 1998). Despite relatively small population numbers, foundresses are readily identifiable early in the season by their larger size and nest-searching behavior, in which individuals fly low to the ground in tight-turning movements while investigating potential nesting burrows (Svensson et al. 2000, O’Connor et al. 2017, Pugesek and Crone 2022). The selection of a nest by a foundress establishes the location from which subsequent foraging movements extend and may profoundly influence the success of her colony later in the season (Suzuki et al. 2009, Pugesek and Crone 2021). As such, a clear understanding of the timing, scale, and site-selection process in foundresses is critical for understanding bumble bee population dynamics and informing patterns observed in studies of worker movement.

When to Move and Why?

Initial movements by gynes are likely related to the search for suitable mates and overwintering sites. Although there may be species-specific preferences for overwintering substrate (Sladen 1912), there is no empirical evidence suggesting gynes are motivated to move long distances before hibernating. Despite observations being limited in geographic scope and species diversity, two recent studies suggest overwintering occurs nearby or even within the natal nest. Pugesek et al. (2023) found large numbers of overwintering B. impatiens queens in their hibernacula within a few meters of where active nests had been earlier in the year. Similarly, Boone et al’s (2022) lone observation of an overwintering B. affinis gyne was within the natal nest itself. Although there may be reasons to disperse if predator, parasite, or pathogen pressure is high at the natal nest site or if overwintering substrate is locally unavailable, we propose four main reasons why the most adaptive behavior for gynes should be to mate and hibernate quickly.

• 1) Resources during the late-summer mating period and spring foraging period may not be spatially correlated, in which case, dispersing to locate high-resource areas before overwintering would provide no consistent fitness advantage.

• 2) If dispersal is density dependent, dispersing prior to overwintering may be maladaptive because it would incur all the risks associated with dispersal, but with less information on the subsequent strength of food or nest competition the following season. Because many queens will die during diapause (Beekman et al. 1998, Pugesek et al. 2023), dispersal may be unnecessary, and the remaining queens will be in local habitats that were successful for their mothers.

• 3) Increased time spent foraging before overwintering incurs additional risks of wing wear, predation, and depleted fat reserves.

• 4) High resource status would lead to more successful overwintering. Gynes have been observed to remain in the colony or return to their natal colony for some time after eclosion (Allen et al. 1978). If reserves are available from nestmates or within the nest, gynes should feed within the colony and avoid foraging alone in the open.

Additionally, males seem to exhibit behaviors aimed at avoiding inbreeding (eg Darvill et al. 2007, detailed below), and so we have little reason to believe avoidance of inbreeding would be motivation enough for gynes to disperse. Furthermore, work on insect physiology suggests many organisms tend to be in one of two contrasting states: a nutrient sequestration state where nutrients are shunted into storage tissues and little flight occurs, and an alternate state where individuals are highly mobile and using their reserves for flight (Boggs 1981, 1992, Zera and Denno 1997, Zera and Harshman 2001). Our knowledge of gyne physiology suggests they are in the former state and ostensibly less dispersive (Alford 1969, Röseler and Röseler 1986). Although dispersal during the gyne phase may occur, it likely only occurs incidentally (eg while looking for suitable overwintering sites; Van Dyck and Baguette 2005) or under circumstances of stress.

Although we posit that gynes likely have little directed dispersal, a foundress’ motivation to choose between dispersal and philopatry following diapause is more compelling. Foundresses need to forage and locate a suitable nest site in landscapes that offer reliable resources (Svensson et al. 2000, Kells and Goulson 2003, Suzuki et al. 2007, 2009, Pugesek and Crone 2022). Decisions to disperse or settle are likely to be determined by the balance between innate individual development and flight behavior, internal nutritional state, and quality of the local environment for forage and nesting resources. Foundresses experiencing nutritional stress will have limited capacity to disperse and would benefit from allocating more time and effort toward foraging. In contrast, individuals not experiencing nutritional stress can afford to allocate more time toward searching for suitable nest sites before settling. Thus, individuals with access to more abundant food resources can also be choosier in their searching patterns. Similar nutritionally-dependent choice scenarios are common across animal systems (Raubenheimer et al. 2009).

In addition to their nutritional status upon emergence, foundresses likely choose to settle or disperse based on their assessments of current forage, soil conditions, nest site availability, or the density of competitors for these resources. Foundress queens emerging in environments with abundant pollen and nectar resources would benefit less from dispersal if suitable nest sites also are available nearby because they would save time and energy for successful nest initiation. However, to ensure nest establishment in a high-quality habitat, it seems likely that some exploration of the surrounding area probably occur before settling on a nest site, regardless of the abundance of resources near the emergence site (Pugesek and Crone 2022). Because foundresses emerge with no knowledge of the present resource landscape, an innate exploratory behavior would allow foundresses a means to sample forage availability and provide the cues to settle and search for nest sites in high-quality habitats (ie even foundresses emerging in high-quality sites should explore before settling, because they will have no prior knowledge that they are in the ‘best’ site).

Regardless, it is known that foundresses occasionally disperse from sites with readily available floral resources, as evidenced by mark-recapture studies of Bowers (1985) and our own observations in California (Mola et al. 2020b), and so other limiting factors such as nest site or resource competition must be considered. More importantly, in these studies, there was no understanding of the internal physiological status of the queens. Dynamic state variable models have allowed for a deeper understanding of the relationship between internal and external factors for insect parasitoids (Collier 1995, Clark and Mangel 2000, Bernstein and Jervis 2008). Similarly, modeling approaches that incorporate both the physiological condition of individuals alongside ecological measurements (eg dynamic state variable models or stochastic individual-based models) could be particularly useful for understanding time allocation and explaining the movements of foundresses.

Post dispersal and after a foundress has chosen a suitable nest site, she must continue to forage for resources to provision her early cohorts of offspring and to maintain her metabolic needs during incubation (Sladen 1912). After the establishment of a strong worker force, the foundress will cease most movements outside the nest and assume a predominantly reproductive, and relatively sedentary, role within the nest (Fig. 1).

Movement Propensity, Potential, and Behavior

As hypothesized above, the realized movements of gynes are likely to be relatively small; however, the movement potential of gynes, should be great given their relatively large body size, greater nutrient stores, and lack of a central nest to return to and tend (Fig. 2). Given the cryptic nature of foraging and dispersal during this phase, little inference from the literature is possible beyond relationships of body size and expected behavior. Research into the movement behaviors of gynes could help us understand whether late-season foraging resources provide greater overwintering success (Woodard et al. 2019) or aid in connectivity and gene flow between populations if such movement occurs during this phase.

The detailed distances and directionality of foundress queens between emergence from diapause and nesting are unknown, but movement potentially can be generalized as a mix of dispersal flights over large areas and short distance flights for foraging and nest searching during which foundresses sample local habitat quality (Pugesek and Crone 2022). A behavioral model for amphibian movement (Pittman et al. 2014) may be useful for considering movement at this stage. Individuals decide to disperse and enter an ‘away’ state followed by ‘directed’ and ‘settlement’ state (Fig. 3). In the away state, individuals should move rapidly in a random (or guided; for example, if there is a trend to follow wind currents or major habitat features such as rivers) direction, resulting in high initial net displacement. Initially, individuals may pass over suitable habitat, but as they transition through a directed phase and amass prior information, their responsiveness to habitat should increase, inducing settlement. The amount of time spent in the directed dispersal mode could be determined by any one, or a combination of, initial cues to move such as conspecific density, lack of floral resources, or the internal physiological state of the individual. For example, individuals can forage for nectar to offset the metabolic cost of long-distance flight, thus those with greater energy stores might fly for longer once they enter a dispersal state because metabolic signals of low energy do not interrupt this behavioral state. The same pattern of displacement could occur without distinct behavioral states. If individuals simply move more rapidly through lower-quality habitats as forage or nest sites are less frequent, large movements could occur until the individual reaches a suitable habitat again (e.g., Pugesek and Crone 2022).

Fig. 3.

Behavioral phase diagram for foundress queens, modified from Pittman et al. (2014). Foundress queens begin their movement with a decision or cue to leave (left) and move through away and directed stages ultimately towards settlement (right). Scale of movement decreases from left to right with net displacement and movement rate declining as foundresses become more focused on settlement, with their habitat responsiveness increasing. In this model, orientation begins in a random direction in the absence of information and gradually becomes biased as responsiveness to habitat increases. Overall, the movement of foundresses is characterized initially as broad scale and coarse-grained moving towards fine scale with high responsiveness as information is gained from the habitat or as nutritional state drives flight motivation and behavior.

This proposed behavioral model is congruent with recent empirical work. Using commercially reared queens following overwintering, Makinson et al. (2019) found that although B. terrestris foundresses spent much of their time resting, they made frequent short movements (mean ± se: 34.2 ± 2.3 m) with limited directionality resulting in long-distance dispersal (>3 km; from modeled movement) over typical pre-nesting flight periods (Makinson et al. 2019). Similarly, Cavigliasso et al. (2020) found that queens B. pauloensis initially moved over large distances prior to nest establishment before settling and foraging over a more restricted area. This process of shifting between long-distance ‘away’ flights and focused ‘settlement’ search patterns is consistent with analogous work for flower-foraging workers showing an encounter with a low-reward flower triggers a long-distance away flight before settling back to a restricted area search (Schulke and Waser 2001). Support more broadly for directed dispersal is documented across a range of insect orders, including butterflies, where movement patterns demonstrate individuals are in an ‘away’ state distinct from displacements that occur as a secondary consequence of other routine movements like foraging (Van Dyck and Baguette 2005). Despite these bits of evidence, much remains to be learned about what induces queens to move or settle. This knowledge gap warrants attention because this information could be critical in designing restored areas in ways that allow them to be linked or detected as suitable habitats (Schultz 1998).

Realized Scale of Movement

The scale of gyne movement is also unknown, but recent molecular advances combined with careful repeated sampling could offer a means for estimating it. Genetic samples taken from foraging gynes at the end of the season, and previously from worker-siblings earlier in the same season could be used to estimate pre-wintering dispersal distances of gynes from their natal nests. In the absence of these data, most estimates are either of combined gyne and foundress movements or foundress dispersal alone. Despite this difficulty, there is a practical reason to determine the distinctions. Long dispersal of both gyne and foundress phases would make queens (and populations in general) more robust to habitat fragmentation, loss of suitable habitat, or overcrowding with conspecifics because individuals could potentially disperse greater distances and follow cues at multiple ontogenetic phases. Such multi-stage movement would be facilitated by renewed resources in different seasons, such as spring trees and ephemeral wildflowers (Fye and Medler 1954, Motten 1986, Mola et al. 2021).

A few studies have attempted to estimate the distances foundresses are capable of dispersing. Mark-recapture estimates of Bombus spp. among mountain meadows indicate that foundresses fly as far as 1.2 km early in the season (Bowers 1985). In our own mark-recapture work in mid-elevation mountain meadows surrounded by chaparral, marked queens of B. vosnesenskii were reobserved at distances up to 3.3 km (Mola et al. 2020b). These methods do not identify the motivation of queens to disperse these distances (ie for forage, nesting sites, or otherwise), but confirm that substantial movement can occur in the foundress phase.

A few other estimates of queen movement exist, but the authors either were uncertain whether they were observing gynes or foundresses (Hagen et al. 2011) or the aggregate of gyne and foundress movements were reported by measuring foundress separation from her natal colony (Lepais et al. 2010, Carvell et al. 2017). Distances of 3 and 5 km for B. pascuorum and B. lapidarius were documented in European agricultural settings (Lepais et al. 2010). Similarly, Carvell et al. (2017) found a mean dispersal distance of 1.23 km for three species (B. terrestris, B. pascuorum, B. lapidarius; see Table 1 for species means), representing values two to three times greater than estimated foraging distances for workers in the same landscape (Redhead et al. 2016). Linear distances of 1.3 km have also been documented for B. hortorum equipped with radio transmitters (Hagen et al. 2011). However, it’s also possible that these estimates are quite low compared to the total capacity for queen movement. Although rates of movement during species invasions may be facilitated by human activity or release from limitations within a species’ home range, dispersal rates by range-expanding B. ruderatus and B. terrestris in South America suggest dispersal of 15 to 95 km per year are possible (Morales et al. 2013). In total, common dispersal distances of queens are probably in the range of a few kilometers, but even just a few long-distance movements can help maintain the panmictic populations documented for several species (Cameron et al. 2011, 2011, Jha and Kremen 2013a, Mola et al. 2024).

Table 1.

Foraging range and dispersal estimates, by species and caste. Caution is taken in interpreting inter-study comparisons due to differences in methodologies among studies (see Mola and Williams 2019 for further discussion), but this table is provided for reference

Species	Mean of mean estimates (m)1	Maximum reported distance (m)2	Reference(s)
Workers
B. balteatus	85.4	541	(Geib et al. 2015)
B. bifarius	67.78	541	(Geib et al. 2015, Mola et al. 2020a)
B. distinguendus	391	955	(Charman et al. 20103)
B. diversus	–	723	(Nagamitsu et al. 20124)
B. flavifrons	23.8	505	(Geib et al. 2015)
B. hortorum	273	810	(Redhead et al. 2016)
B. humilis	475	2,530	(Connop et al. 2011)
B. hypocrita	–	848	(Nagamitsu et al. 20124)
B. lapidarius	517	1,260	(Walther-Hellwig and Frankl 20005, Knight et al. 20054, Carvell et al. 2012, Redhead et al. 2016)
B. muscorum	100	500	(Walther-Hellwig and Frankl 20005)
B. pascuorum	474.67	1,082	(Darvill et al. 20045, Knight et al. 20054, Carvell et al. 2012, Redhead et al. 2016)
B. pratorum	–	674	(Knight et al. 20054)
B. ruderatus	502	2,125	(Hagen et al. 2011, Redhead et al. 2016)
B. sylvarum	231	2,280	(Connop et al. 2011)
B. sylvicola	74.7	681	(Geib et al. 2015)
B. terrestris	528.5	1,669.5	(Osborne et al. 1999, 2008, Walther-Hellwig and Frankl 20005, Darvill et al. 2004, Knight et al. 20054, Hagen et al. 2011, Redhead et al. 2016)
B. vosnesenskii	1,278	11,600	(Rao and Strange 2012, Jha and Kremen 2013b, Mola et al. 2020a, 2020b)
Queens
B. lapidarius	–	5,000	(Lepais et al. 2010)
B. pascuorum	–	2,074.5	(Lepais et al. 2010, Carvell et al. 2017)
B. terrestris	–	1,553	(Carvell et al. 2017)
B. hortorum	–	1,300	(Hagen et al. 20114)
B. vosnesenskii	2,116	8,103	(Mola et al. 2020b)
Males
B. terrestris	–	2,600–9,900	(Kraus et al. 2009)

¹The mean of the mean estimate in each study. Some species are only represented by a single study, so the mean is just the one estimate. In some instances, a mean estimate was not provided by the authors. We note how we handle these references in further footnotes.

²Typically, the largest reobservation or sibling separation difference, exceptions noted.

³The authors provided the mode, rather than mean, in this study. Additionally, the maximum range is that in which 95% of all colony activity occurred.

⁴Only maximum estimate reported, no mean.

⁵Mean is the foraging radius where most activity occurred, maximum is the largest observed radius.

Shifting from dispersal to foraging, theoretical considerations in a foundress’ life suggest that she should forage closer and for less time than her maximum capability (Fig. 2), both for efficient resource return and to maintain consistent incubation temperatures of her brood (Goulson 2009). Time devoted to foraging likely also reduces the number of eggs a foundress can successfully tend, making the incentive to quickly produce the first cohort much greater (Sarro et al. 2021). Preliminary evidence seems to support these contentions, suggesting that foundresses make frequent, shorter duration foraging bouts (Gustilo et al. unpublished manuscript) compared to durations from other studies on workers. However, further study with larger sample sizes and within-study comparisons of castes is needed to determine the consistency of this pattern. Such information could inform conservation management schemes by determining if foundress foraging ranges are in fact shorter and less plastic than those of workers, suggesting the benefit of plentiful resources near nesting habitat. This would further help to confirm observations that between-year colony reproduction and survivorship appear to be particularly sensitive to the abundance of spring resources (Carvell et al. 2017).

We conclude that the movement potential of queens should be relatively large compared to workers due to the queen’s greater nutritional stores, associated larger body sizes, and multiple opportunities to disperse between eclosing and nesting. Evidence from limited studies on invasions (Morales et al. 2013) and subsequent range expansions or potential ‘migratory’ behavior supports this perspective (Fijen 2021). However, their realized movements may typically be much less than their potential due to risk-reward trade-offs in dispersing from the natal site, and time trade-offs with incubation or other demands during colony establishment (Fig. 2). Little is known about gyne movements or the effects of habitat corridors, barriers, or landscape heterogeneity on queen dispersal. Additionally, much of what we do know about the queen movement is from a limited number of studies, species, and contexts. This knowledge gap presents one of the most crucial challenges for understanding the bumble bee movement and informing conservation efforts. Overall, new research into queen dispersal and foraging will help to inform species conservation efforts by providing information on when, where, and how habitats and resources are connected through space and time.

Males

Introduction

The movement of male bumblebees has distinct consequences for bumble bee populations and associated pollination services. Unlike non-reproductive workers and reproductive (ie nesting) queens, males are not restricted to central-place foraging. Instead, once a male leaves his natal nest, he does not return, potentially dispersing great distances (Kraus et al. 2009, Wolf et al. 2012). As such, the movements of males could play a large role in the genetic connectivity of bumblebee populations, and potentially in the long-distance transfer of pollen among plants they visit (Ogilvie and Thomson 2015). However, our understanding of the scale, pattern, and frequency of male movements is largely unexplored. We review the few studies that have investigated male dispersal and speculate on further movement patterns from studies of mating behavior and theoretical considerations.

When to Move and Why?

Males are typically produced before queens (Fig. 1; Pomeroy and Plowright 1982, Bourke 1997, Beekman and van Stratum 1998, Beekman et al. 1998), and should disperse from the natal nest to minimize the chances of inbreeding. Indeed, they appear to have at least some ability to avoid sib-mating by aggregating outside nests of non-sibling (Darvill et al. 2007).

Movement Propensity, Potential, and Behavior

Males move to acquire resources to fuel their own metabolism and to locate mates. These behaviors should result in unintentional or routine dispersal, likely impacting overall displacement (Van Dyck and Baguette 2005). Combining the need to avoid sib-mating and to maximize mate location, we hypothesize that males might first show a brief period of directed dispersal, and then search over a smaller area using a variety of species-specific mate-searching behaviors (Fig. 2; Villalobos and Shelly 1987).

Most of what is known about male movement is from studies of mate-searching. It is unknown if these mating behaviors contribute to dispersal or large-scale movements, either directly or incidentally. For instance, male bumble bees, like many other insect taxa, exhibit hilltopping behaviors, gathering at high points to locate mates (Goulson et al. 2011). Depending on the topography of the landscape and the distances traveled between natal colonies and hilltops, hilltopping could impact the dispersal of males (eg such as theorized and documented for Lepidoptera; Painter 2014, Grof-Tisza et al. 2017). Darwin (1886) and others documented males exhibiting trapline-like behavior wherein scent-marked paths were routinely visited and traveled (Freeman 1968, Alcock and Alcock 1983, Bergman and Bergström 1997). Although observations of queens or mating events along such flight traplines are rare, they do provide another example of an evident and widespread male movement propensity. Bumble bee species differ in the morphology of males, notably that of the eyes. Some species have enlarged eyes and are known to perch and look for mates from a distance (Franklin 1954, O’Neill et al. 1991, Streinzer and Spaethe 2014). Others have eyes of more similar size to workers and patrol routes looking for mates. Within and among species, there is also considerable variation in body size amongst males (Amin et al. 2012). Whether these morphological differences correlate with a tendency to disperse, however, is unknown.

Realized Scale of Movement

The two primary studies of the movement scale of male bumble bees (Kraus et al. 2009, Wolf et al. 2012) infer male flight distance indirectly using molecular techniques, and assumptions about the foraging area of the worker castes within their sampling region. In both studies, the authors collected male and worker bumble bees within a common sampling area. They then assigned all individuals to queen genotypes using sibship reconstructions (Darvill et al. 2004, Wang 2004) and conclude that any males assigned to queen genotypes undetected in their worker collections must have flown from outside the foraging area of the workers, to within the sampling area. As such, the minimum estimate of male flight distance is the radius of the worker foraging area (Kraus et al. 2009). Kraus et al. (2009) concluded that the male flight distance for B. terrestris is 2.6 to 9.9 km. In Wolfe et al. (2012) they calculate the relevant area to be 1.66 and 1.74 times that of worker foraging ranges for B. terrestris and B. lapidarius, respectively. Although there are certain limitations to these studies, notably limited marker numbers, lumping of non-detections and novel detections, and the lack of actual mark recapture of males, the estimated distances align well with estimates of workers and queens of these species (eg Redhead et al. 2016, Carvell et al. 2017). Additionally, the results parallel conceptual considerations of the relationship between scale across castes due to the multiple possible movements of males (Fig. 2).

Overall, research on male dispersal or foraging, and its consequences for bumble bee population demography and genetic connectivity is poorly resolved. Efforts to improve our understanding in this realm could be well rewarded, as we have conceptual reasons to believe male dispersal, given their lack of restriction to central-place foraging, may be critical in establishing long-range gene flow among populations.

Workers

Introduction

The movement of worker bumble bees is distinct from that of males and queens. Workers are central-place foragers and so do not disperse. As such they are not capable of directly affecting immigration, emigration, and colonization. However, their foraging movements are fundamental to colony success and reproduction and thus indirectly impact population dynamics. Because their resource acquisition affects the nutritional status of eclosing gynes, their movements also likely influence population genetic structure and potential population connectivity. Thus, understanding their movement is critical to properly designing conservation schemes not to mention planning pollination services.

When to Move and Why?

Workers internal motivation to move for foraging is driven by the demand of a growing larval population, colony resource status, and forager recruitment (Cartar and Dill 1990, Cartar 1992, Plowright et al. 1993, Dornhaus and Chittka 2001). In young colonies, the amount of pollen collected by the worker population is proportional to the size of the larval population (Pendrel 1977). Experimental removal of pollen or nectar reserves from the colony increases foraging activity (Cartar 1992, Plowright et al. 1993) whereas addition reduces it. Additionally, the needs for pollen and nectar do not directly correlate over time, but rather the colony responds to the need for each resource individually (Plowright et al. 1993). The close relation between foraging activity and resource levels or demand from within the colony further indicates that workers will avoid the risks associated with foraging rather than indefinitely building the colony’s supplies (Pelletier and McNeil 2004, although see Koch and Cane 2022 for an example of potential overcollection of resources). Although bumble bees do not appear to share social information about resource location, workers do monitor resource supplies within the colony and communicate the need for additional foraging (Dornhaus and Chittka 2001) and in some species, alert their nestmates when food resources are discovered (Dornhaus and Cameron 2003, Dornhaus and Chittka 2004).

Worker movement changes not only based on resource demand and motivation by other colony members, but also with ontogeny of behaviors over their lifetime. When individuals first eclose, they stay within the colony working on nest-related tasks such as brood care or hygienic behaviors (Sladen 1912). After the first few days, they begin to leave the nest to collect food resources, although some individuals, typically smaller workers, never leave the nest. Beyond the initial days after eclosion, there appears to be no association between age and task performance (Cameron 1989). New foragers conduct a series of orientation flights presumably to ensure they can relocate the nest, but also to scout potential forage patches (Osborne et al. 2013, Woodgate et al. 2016). Some individuals maintain exploratory behaviors and attraction to novel patches (Woodgate et al. 2016) or plant species (Heinrich 1979), but it is unclear whether this is explained largely by innate individual differences in behavior or changes in either individual or colony development over the season. Many workers then express strong fidelity to specific flower patches, even as resources change either in their quality (Thomson 1988) or to entirely different species (Ogilvie and Thomson 2016).

Movement Propensity, Potential, and Behavior

Empirical studies reveal considerable plasticity in bumblebee foraging ranges and behaviors (Jha and Kremen 2013b). Our ability to generalize results across studies is limited by differences in sampling methodology, landscape context, study species, and scale of estimation (Mola and Williams 2019). Although clear trends exist in the relative rank of species movement distance (Mola and Williams 2019), targeted research could reveal when these differences are due to species traits, ontogeny, nutritional status, landscape response, or various other factors that may alter movement patterns and scale. Theoretical models of bumble bee movement provide reasoned predictions from which to consider deviations that may result from various internal and external factors.

Theoretical models have proven useful for inferring the potential scale of movement for worker bumble bees. Cresswell et al. (2000) modeled the foraging range based on the energetics of foraging, balancing the cost of flight with the resource return rate and quality of resources in the landscape. Compiling parameter values from various studies, they predict bumble bee flight is economically viable up to several kilometers (Cresswell et al. 2000). Similarly, Dukas and Edelstein-Keshet (1998) predicted a maximum viable flight range of 5.4 km for bumble bees.

In addition to energetics-based models of foraging, empirical studies suggest potential flight distances of workers may sometimes be great. For example, homing studies have documented successful returns at distances up to 9 km (Goulson and Stout 2001). These studies do not speak to how commonly individuals forage at these distances, nor whether trips at this range are economically profitable for the colony. They do suggest a very long potential flight capacity and strong spatial-cognition abilities allowing individuals to relocate their nests, likely by making use of landmarks and ground-level features (Brebner et al. 2021).

Realized Scale of Movement

The scale over which foragers move has been critically debated for decades. It was originally believed that bumble bees foraged over relatively short ranges (Heinrich 1976, Kevan and Baker 1983, Bowers 1985). However, this inference was likely due to the difficulty of locating individuals away from their nest because the area over which researchers must search increases exponentially with distance from the nest, and combined with the site fidelity of foragers (Dramstad 1996, Ogilvie and Thomson 2016, Mola and Williams 2019) finding such a long-distance forager is probabilistically low. Contemporary work suggests that individuals and colonies operate over larger scales, commonly ranging from about 300 to 1500 m, depending on the study species, landscape, and sampling methodology (Table 1).

The foraging range should be highly plastic so that workers and colonies can cope with phenological turnover in plant availability in patchy landscapes during the flight season (Jha and Kremen 2013b, Pope and Jha 2018, Mola et al. 2020a, Hemberger and Williams 2025). Furthermore, it appears both conceptually and with some empirical evidence that foraging range scales positively with species-specific colony sizes (Westphal et al. 2006; although this could partly reflect the increased probability of sampling species with larger colonies at greater distances). We extend this reasoning to within-species seasonality, suggesting that the foraging area should expand as the colony grows and resource demands increase. Recent empirical evidence also demonstrates that the foraging range extends as resource quality declines over the season, with foraging occurring over greater distances to reach higher-quality patches (Pope and Jha 2018). Conversely, this suggests that foraging ranges generally should be smaller in species with short flight seasons, and in landscapes where the flowering season is brief (eg Geib et al. 2015).

Foraging distances of individuals and colonies appear to be dynamically driven by changing resource availability and composition. Jha and Kremen (2013b) found that B. vosnsenskii foragers travel greater distances to reach species-rich patches, suggesting that diversity, rather than only abundance, affects foraging range. This point is supported by work documenting the tendency of some Bombus species to select mixtures of flower species that meet specific nutrient ratios (Vaudo et al. 2015, 2016) and evidence suggesting workers from the same colonies will forage greater distances to mix resources as plant communities turn over with increased spatial separation (Mola et al. 2020a). Carvell et al. (2012) found that the colony-specific foraging range of B. lapidarius and B. pascuorum decreased with an increasing proportion of forage habitats in the surrounding landscape. This pattern suggests that foraging efforts or recruitment to distant patches declines as a greater abundance of resources are available locally. Similarly, Hemberger and Gratton (2018) found that the foraging trip duration of B. impatiens declined with a resource-pulse in the landscape, further suggesting the dynamic nature of bumble bee foraging. These studies taken together suggest that although bumble bees are capable of long-distance movements, the realized distances foragers travel are highly context specific, with distances modified depending on the richness, abundance, density, and phenology of resources.

Overall, the foraging movements of workers are highly context specific, and estimates should be interpreted with awareness in their ecological and methodological context rather than as fixed values. Intra- and inter-specific studies are needed to reveal whether differences observed thus far are due to species-specific traits, landscape context, or methodological constraints. Of note is the lack of study on how changes in foraging distance influence the reproductive output of colonies, similar to studies for solitary bees (eg Zurbuchen et al. 2010). Future studies making use of dynamic measurements of bumble bee foraging, specifically linked to reproductive output, will give us a clearer understanding of how changes in phenology or land use may affect bumble bee foraging patterns and allow us to parameterize connectivity models, effective design restoration plantings, or inform the placement of colonies for crop pollination.

Modifiers of Bumble Bee Movement

Above, we have discussed bumble bee movement motivations, capacity, and realized scale of movements relatively agnostic to external and internal factors that may modify the likelihood or extent of foraging and dispersal. In real ecological systems, these external (eg landscape structure, floral diversity, patch size) and internal (eg pesticide exposure, infection status) factors certainly affect movement (Ricketts 2001, González et al. 2024). Understanding how these different internal and external factors modify movement is a likely frontier of insect movement ecology, especially given the rate and magnitude of global change drivers. Determining how the distance or probability of foraging and dispersal are affected by, for example, landscape structure, composition, and fragmentation as well as pesticide exposure, changing climate, or shifting phenology are all fruitful areas of research.

Inversely, we may also be interested in using movement diagnostically. Instead of asking how these factors change movement, we may ask what observed changes in movement reveal about landscapes or bumble bee fitness. For example, changes in the speed, distance, or directionality of movement in response to varied levels of pesticide exposure across multiple species could reveal which species are more sensitive to these types of stressors. Similarly, differences in movement characteristics have been successfully used to assess the habitat preferences of foundresses for nesting (Pugesek and Crone 2022). In that case, the characteristics of flight paths and frequency of stereotyped behavioral modes changed with habitat type as an indicator of preferred nesting. Identifying stereotypical behaviors of bumble bee movement and standardizing methodology to assess these characteristics could yield reliable and simple-to-use protocols for assessing the success of management interventions across a broad range of species and ecosystems.

External Modifiers of Movement

How external factors like landscape connectivity, structure, and patch arrangement affect bumble bee movement is of great practical interest in landscape planning and assessing the effects of landcover change. Over broad scales, urbanization (Jha and Kremen 2013a) and oceanic barriers (Darvill et al. 2010, Jha 2015) limit bumble bee dispersal resulting in population differentiation. These disruptions of functional connectivity at scales of tens of kilometers likely result from a combination of reduced habitat suitability and impeded queen dispersal but are not necessarily surprising over these distances. At scales of more immediate management concern, representing the dispersal or foraging distances of individuals (ie <10 km or so), somewhat contrasting patterns exist. Bumble bees seem relatively robust to the effects of putative barriers like forests or changing landcover. Experimental colonies placed in forest fragments found colonies in ‘isolated’ patches performed no differently than colonies placed in putatively connected patches (Herrmann et al. 2017). Similarly, studies demonstrate individuals and colonies readily cross forest boundaries to forage (Kreyer et al. 2004, Mola et al. 2020a). Despite this resilience to putative barriers, bumble bees also demonstrate clear behavioral responses to edges, linear features, or hard transitions in landcover. For example, although foragers will cross features like roads if experimentally relocated, they seem resistant to do so independently (Bhattacharya et al. 2003). Additionally, radio tracking of individuals makes it clear that they use linear features to navigate to resources and the presence of linear features in a landscape influences their exploratory behaviors and navigation patterns (Brebner et al. 2021). Overall, it seems that bumble bees use features like forest edges, roads, and rivers to aid in navigation and learning, but they do not present strong barriers to their movement. Instead, movement seems most strongly influenced by a response to floral abundance, diversity, and phenology, regardless of the interceding landcover (Jha and Kremen 2013b, Pope and Jha 2018, Mola et al. 2020a, Dániel-Ferreira et al. 2022).

Internal Modifiers of Movement

Beyond external modifiers, internal factors like pesticide exposure, pathogen load, nutritional status, and age or experience also modify potential and realized patterns of bumble bee movement. There is clear experimental evidence from lab studies that pesticide exposure negatively impacts flower handling, spatial foraging behavior, and flight capacity of foragers (Gill et al. 2012, Samuelson et al. 2016, Kenna et al. 2019). Exposure to extreme hot or cold temperatures similarly limits flight endurance (Kenna et al. 2021). How these lab studies scale to field-realistic scenarios is a critical area for future research. Other internal factors that on principle ought to strongly influence movement are surprisingly poorly understood. Despite clear evidence of widespread pathogen exposure in wild bumble bees (Schmid-Hempel and Durrer 1991, Plischuk et al. 2009, Li et al. 2012, Gillespie and Adler 2013), how pathogen infection impacts movement, subsequent resource gain, and population connectivity is less resolved. Similarly, internal nutritional status (Woodard and Jha 2017) ought to have a large impact on the timing and extent of foundress movements as well as on the realized movement patterns and distances of individual workers and colonies. In our review of the movement potential of life stages, we discussed several movements that are likely either induced or diminished by nutritional status. For example, gynes with high nutrient levels during post-eclosion nutrient sequestration probably do not forage much before overwintering, whereas individuals that have not had access to high-quality food resources prior to leaving the natal nest likely do. Several aspects of bumble bee movement ecology are also likely to be strongly influenced by colony-level nutritional status. However, the effects of nutritional status remain largely unresolved in either lab or field scenarios. Because nutritional state likely plays a large role in determining flight motivation and capacity, research into the relationship between nutritional status and movement can reveal insights into the causes and consequences of long-distance movements for bumble bees.

Future Studies in Bumble Bee Movement Ecology

Our review highlights the groundswell of research related to bumble bee movement which has greatly advanced our understanding of how bumble bees navigate, perceive, and interact with landscapes. It also reveals several areas of bumble bee movement that are poorly understood or have scarcely been examined. Some of these are especially urgent as they could inform the timing or spatial extent of conservation interventions. Substantial uncertainty remains in understanding dispersive movements that occur at multiple stages of the bumble bee life cycle (Fig. 1). In addition, there remains much to learn about how ecological context, environmental drivers, and nutrition influence movement. Our review clarifies the need for an ontogenetic perspective towards bumble bee movement. However, the more focused study is needed that explicitly considers ecological context (external factors) and its interaction with nutritional and other physiological status (internal factors), minimizes inferences influenced by the selection of study method (Mola and Williams 2019), and to expand our coverage of species and geographic areas.

Priorities for future study may include understanding the propensity of gynes to disperse prior to overwintering. This could be achieved using experimental colonies, of a range of species, in conjunction with harmonic radar, or radio telemetry to track dispersion following departure from the natal nest. This would inform whether species are relatively plastic in their overwintering locations or if late-season resources are important foraging opportunities for queens prior to overwintering, and the influence of their proximity to the natal nest. Further studies are needed to develop more generalized comparisons among species and geographic areas. This need for greater coverage is not just for obtaining greater generality for generality sake, but is critical in conservation planning as we already have some evidence that dispersal propensity explains genetic structuring and sensitivity to fragmentation in some species (Darvill et al. 2010), and there are consistent species-specific differences in foraging range (Mola and Williams 2019).

Conclusions

Bumble bees have been valuable model organisms for fine-scale studies of forager behavior and a growing number of studies seek to investigate landscape-scale movements. Our approach to reviewing movement by treating each caste and life stage in turn highlights the importance of understanding the internal motivations, movement potential, and motion capacities of each caste or behavioral phase. For example, although the larger size of queens suggests an increased potential foraging range relative to workers, their realized foraging movements may be much shorter because of their need to incubate brood and guard the nascent nest. Without such a stage-specific perspective, we might wrongly apply estimates from one stage to another. Forgoing an ontogenetic approach to movement ecology would be akin to assuming all bumble bee species are similar in their movement capacity, which we know to be false. Indeed, the review brings to light gaps in our knowledge of landscape-scale bumble bee movements. Primary among these limitations are studies with queens and males. Specifically, if our knowledge of worker movements and responses to external and internal modifiers of movement translates to those of the reproductive castes. Perhaps of most apparent use in the immediate term is the recognition of stereotypical movement behaviors that allow us to use observed bumble bee movement diagnostically to reveal other aspects of landscapes or bumble bee life history of interest (eg Pugesek and Crone 2022).

Our extension of bumble bees as a model system for studies of movement beyond foraging highlights the need for dynamic, ontogenetic considerations within movement ecology research generally. The development of colonies, differences between castes, or individual body size, has parallels in almost all mobile organisms, with ontogenetic and morphological differences being pervasive for understanding movement generally. As individuals shift among states varying in their mobility or resource demands, the scale over which conservation action should be considered changes as well. Hopefully, our review provides guiding principles for predicting and interpreting bumble bee movement within the context of castes or life-stage-based considerations.

Author contributions

John Mola (Conceptualization [equal], Writing—original draft [lead], Writing—review & editing [equal]), and Neal Williams (Conceptualization [equal], Writing—review & editing [equal])