Variation in North American bumble bee nest success and colony sizes under captive rearing conditions

Abstract

Of the 265 known bumble bee (Bombus) species, knowledge of colony lifecycle is derived from relatively few species. As interest in Bombus commercialization and conservation grows, it is becoming increasingly important to understand colony growth dynamics across a variety of species since variation exists in nest success, colony growth, and reproductive output. In this study, we reported successful nest initiation and establishment rates of colonies and generated a timeline of colony development for 15 western North American Bombus species, which were captively reared from wild-caught gynes from 2009 to 2019. Additionally, we assessed variation in colony size among 5 western North American Bombus species from 2015 to 2018. Nest initiation and establishment rates varied greatly among species, ranging from 5–76.1% and 0–54.6%, respectively. Bombus griseocollis had the highest rates of nest success across the 11-yr period, followed by B. occidentalis, B. vosnesenskii, and B. huntii. Furthermore, days to nest initiation and days to nest establishment varied among species, ranging from 8.4 to 27.7 days and 32.7 to 47 days. Colony size also differed significantly among species with B. huntii and B. vosnesenskii producing more worker/drone cells than B. griseocollis, B. occidentalis, and B. vancouverensis. Additionally, gyne production differed significantly among species with B. huntii colonies producing more gynes than B. vosnesenskii. Results from this study increase knowledge of systematic nesting biology for numerous western North American Bombus species under captive rearing conditions, which can further improve rearing techniques available to conservationists and researchers.

Introduction

Bumble bees (Hymenoptera: Apidae: Bombus) are abundant and diverse pollinators with approximately 265 species primarily distributed throughout temperate, alpine, and subarctic ecosystems worldwide (Kremen et al. 2002, Berenbaum et al. 2007, Klein et al. 2007, Goulson 2010, Strange and Tripodi 2019, Williams and Jepsen 2020, Maebe et al. 2021). Bumble bees are primitively eusocial insects with overlapping generations within a colony, reproductive division of labor, and cooperative care of offspring. With the exception of some tropical species (Taylor and Cameron 2003), most bumble bee species have an annual life cycle that begins with the emergence of a mated gyne (reproductive female) from winter dormancy. The gyne searches for a suitable nesting site in abandoned rodent burrows, open grass tussocks, hollow logs, or above-ground man-made structures, and then forages for nest provisions (i.e., pollen and nectar) (Williams et al. 2014). The foundress gyne (now queen) constructs a wax honeypot for nectar storage within the nest, oviposits the first brood clutch on a pollen mass moistened with nectar, incubates the first clutch of brood, and continues to forage to provide food for the larvae (Williams et al. 2014). Each individual brood cell contains a single bumble bee larva. After the emergence of workers (female offspring) from brood cells, the queen ceases foraging. The queen then focuses efforts on oviposition, brood care, and maintaining social order, while workers perform tasks related to foraging, brood care (e.g., feeding developing larvae), colony maintenance (e.g., thermoregulation, cleaning), and defense. As floral resources become more abundant, the queen can produce offspring quickly, often resulting in rapid colony growth and development. Once the colony population peaks, usually in midsummer, the queen produces new reproductive gynes and drones (males). The new gynes and drones then leave the colony to feed and mate with offspring from other colonies. The foundress queen, workers, and drones then die; the newly mated gynes find a subterranean space and create a hibernaculum to undergo winter diapause, and the cycle continues (Alford 1975, Goulson 2010, Strange 2010, Williams et al. 2014, Koch et al. 2021).

Among the diverse Bombus species, there is considerable variation in colony establishment, growth, and development (Plowright and Jay 1966; Alford 1970; Pomeroy and Plowright 1980, Strange 2010). For example, nest site selection varies among species. While most species nest in subterranean habitats, such as abandoned rodent burrows, some others nest on soil surfaces in grass tussocks (Hobbs 1965a, 1965b, 1967, Macfarlane et al. 1994, Taylor and Cameron 2003) or in tree cavities, bird houses, or other elevated cavities (Macfarlane et al. 1994, Koch and Cane 2022). Furthermore, timing of gyne emergence from diapause is known to vary among species. Some Bombus emerge early in the spring, while others are not seen until after the emergence of early spring workers (Colla et al. 2011; Koch et al. 2012; Williams et al. 2014). These differences likely have evolutionary significance and may serve to reduce interspecific competition for nest sites and/or ephemeral spring floral resources, or to avoid social parasitism. In addition to species-specific nest sites and emergence timing, there is variation in colony size, nest success, and reproductive output among species. This variation can also be seen in laboratory-reared colonies, which is important to the utility of Bombus species as commercial pollinators.

Given these biological and evolutionary differences in colony-level factors, it is necessary to determine differences in nest initiation and establishment success, number of offspring and reproductives produced per colony, and colony development timelines among Bombus species (Macfarlane et al. 1994, Kwon et al. 2006, Yoneda 2008, Strange 2010). However, capturing this information in wild colonies is difficult given that the nests are cryptic and often below ground. Obtaining data from colonies reared from wild-caught queens in controlled laboratory environments can increase knowledge on Bombus species nesting biology, although rearing bumble bees for research purposes has resulted in varying degrees of success (Plowright and Jay 1966, Kearns and Thomson 2001, Evans et al. 2007, Salvarrey et al. 2013, Ptáček et al. 2015, Strange 2015, Carnell et al. 2020, Christman et al. 2022).

Recent efforts toward rearing bumble bees in captivity have improved techniques available to conservationists and researchers, which can be used to assess the commercial viability of nondomesticated Bombus species and to establish conservation techniques for imperiled species (Macfarlane et al. 1994, Strange et al. 2011, Christman et al. 2022, Rowe et al. 2023). For example, conservation propagation methods (i.e., assisted reintroductions) have been proposed as a recovery action to augment wild populations (Smith et al. 2020). In this study, we reported nest initiation success and colony establishment rates and generated a timeline of colony development for 15 western North American Bombus species, which were captively reared from wild-caught gynes from 2009 to 2019. Additionally, we assessed variation in colony size among 5 western North American Bombus species from 2015 to 2018. Results from this study increase knowledge of systematic nesting biology under captive-rearing conditions, which can further improve rearing techniques available to conservationists and researchers.

Methods

Bombus Rearing

A total of 3,355 gynes from 19 Bombus species were net-collected while foraging after emerging from winter dormancy in the western United States from 2009 to 2019. The captured gynes were transferred from nets to individual 7 × 3 cm plastic vials (W. W. Grainger Inc., Lake Forest, IL) modified to have ventilation holes (Rowe et al. 2023). The gynes were then transported in chilled, insulated containers to the United States Department of Agriculture–Agricultural Research Service, Pollinating Insect–Biology, Management, and Systematics Research Unit in Logan, UT. Once at the laboratory, colonies were initiated following methodology outlined in Rowe et al. (2023). Briefly, the gynes were moved into 15 cm × 15 cm × 10 cm plastic rearing chambers (Biobest Canada, Leamington, ON, Canada) in a designated rearing space maintained at 28 ± 2 °C and 65 ± 2% relative humidity in complete darkness. In approximately 14% of cases, gynes were paired with conspecifics to increase oviposition and nesting success, which is known as cofounding or pleometrosis (Sladen 1912, Plowright and Jay 1966, Bernasconi and Keller 1996, Ptáček et al. 2000, Strange 2010). Cofounding increases social stress, which generally results in the dominant gyne attacking and killing the other gyne. This then increases oviposition and establishment success rates (Sladen 1912, Plowright and Jay 1966, Bernasconi and Keller 1996, Ptáček et al. 2000, Strange 2010). After placement into rearing chambers, each gyne or pair of gynes was provisioned with 2 g of a multifloral honey bee collected pollen diet mixture and a feeder filled with 50% sugar solution with additives (artificial nectar) (Christman et al. 2022, Rowe et al. 2023). Preparation of multifloral pollen provisions and artificial nectar followed methodology described in Rowe et al. (2023). Following the production of brood, each colony was fed pollen and artificial nectar ad libitum. Once 5 workers eclosed, the nest was transferred to a 29 cm × 22 cm × 13 cm plastic colony box (Biobest Canada, Leamington, ON, Canada) to provide additional space for colony growth and development. Activities that involved colony handling, such as transfer, feeding, maintenance, and data recording, were conducted under red light to reduce disturbance to the colony.

Bombus Nest Success

Nineteen Bombus species were reared over the 11-yr period at varying success rates to provide experimental colonies for scientific studies (Supplementary Table 1). However, 4 of the 19 species were collected at low rates and did not establish brood (B. caliginosus Frison = 1; B. morrisoni Cresson = 8; B. pensylvanicus sonorus De Geer = 1; B. sitkensis Nylander = 5); therefore, they were not included in our analysis. Colonies were checked at least every 3 days over the course of their development. Gynes were given approximately 21 days to produce brood after which time they were considered successful. Gynes that did not produce brood within this time range were culled. For each successful colony, the number of days to first brood and the number of days to first worker eclosion were recorded to determine timing of nest initiation and establishment (Strange 2010) and to generate a timeline of colony development in a controlled laboratory environment. Nest initiation was defined as evidence of the queen to produce brood, and nest establishment was defined as the ability of a queen to rear one adult female offspring (worker) from brood (Strange 2010). The timeline of colony development was documented as the average number of days to first observed brood (nest initiation) from nest installment and the average number of days to first observed worker (nest establishment) from nest installment for each Bombus species. The timeline indicates overall colony averages for days to nest initiation and establishment, not individual larval development.

Bombus Colony Size

Given that these colonies were reared to provide colonies for other scientific studies, there is variation in data collection and documentation. From 2015 to 2018, 233 colonies across 9 species were reared in the controlled laboratory setting for the entirety of their lifecycle. Species were only included when more than 20 colonies were represented within the dataset; therefore, 4 species were excluded from our analysis (B. appositus Cresson = 1; B. californicus Smith = 1; B. centralis Cresson = 2; B. melanopygus Nylander = 1). Once the colony stopped producing brood, indicating the end of the colony’s lifecycle, the colony was dissected to assess variation in colony size among the captive-reared Bombus species. The total number of emerged workers/drones brood cells and emerged gyne brood cells was recorded for each colony. Worker and drone brood cells could not be differentiated due to similarities in cell sizes, so they were documented together. Meanwhile, gyne brood cells are produced toward the end of the colony lifecycle, are approximately twice the size of worker/drone brood cells (e.g., Koch and Cane 2022), and often occur on the top layer of the nest in clumps, which allow the gyne cells to be differentiated from worker/drone cells.

Data Analysis

Two-sample z-tests for proportions were conducted for both nest initiation and nest establishment to compare success rates between colonies reared from individual or paired gynes. Individual analysis of variances (ANOVAs) were used to determine both differences in the total number of eclosed workers/drones and total number of eclosed gynes among species over the 4-yr period (P < 0.05). Tukey’s HSD post hoc tests were used when the ANOVAs produced significant results in order to determine which species means were significantly different. All conditions, including normality, variance, and independence, were met for the individual ANOVAs. Statistical analyses were conducted using base functions in R version 4.0.3 (R Core Team 2021).

Results

Bombus Nest Success

From 2009 to 2019, 41.3% of gynes from 15 species produced brood cells (our criterion for nest initiation) (Table 1) and 18.8% of colonies from 12 species produced at least 1 worker (our criterion for nest establishment) in a controlled laboratory environment. Nest initiation and establishment rates varied greatly among species, ranging from 5% to 76.1% and 0% to 54.6%, respectively. Bombus rufocinctus Cresson (initiation: 5.2%; establishment: 3.4%) and B. nevadensis Cresson (initiation: 10%; establishment: 5%) had the lowest rates of nest success from 2009 to 2019. Meanwhile, Bombus griseocollis De Geer had the highest rates of nest success (initiation: 76.1%; establishment: 54.6%) across the 11-yr period, followed by B. occidentalis Greene (initiation: 59.2%; establishment: 34.8%), B. vosnesenskii Radoszkowski (initiation: 48.2%; establishment: 25.2%), and B. huntii Greene (initiation: 38.8%; establishment: 14.3%). Furthermore, days to nest initiation and days to nest establishment varied among species, ranging from 8.4 to 27.7 days and 32.7 to 47 days, respectively (Fig. 1).

Table 1.

Rearing success of western North American Bombus species as defined by the production of brood (nest initiation) and emergence of a worker (nest establishment) from 2009 to 2019. Colony development of Bombus species within captivity as defined by days to nest initiation ± SD and days to nest establishment ± SD

Bombus species	Subgenus	Years produced	Successful nest initiation	Successful nest establishment	Days to first brood	Days to first worker
B. appositus	Subterraneobombus	2011, 2013, 2015, 2016, 2017, 2019	8/59 (13.6%)	4/59 (6.8%)	19.5 ± 10.7	32.7 ± 12.6
B. californicus	Thoracobombus	2014, 2015, 2016	2/8 (25%)	0/8 (0%)	21.5 ± 20.5	NA
B. centralis	Pyrobombus	2009, 2010, 2011, 2014, 2015, 2016, 2019	9/61 (14.8%)	2/61 (3.3%)	17.1 ± 5.5	40.0 ± 1.4
B. fervidus	Thoracobombus	2015, 2016, 2017	2/9 (22.2%)	0/9 (0%)	18.5 ± 3.5	NA
B. flavifrons	Pyrobombus	2013, 2014, 2015, 2016	1/8 (12.5%)	0/8 (0%)	9.0 ± 0.0	NA
B. griseocollis	Cullumanobombus	2011, 2015, 2016, 2017, 2018, 2019	156/205 (76.1%)	112/205 (54.6%)	8.4 ± 6.7	43.2 ± 13.4
B. huntii	Pyrobombus	2009, 2010, 2011, 2012, 2013, 2014, 2015, 2016, 2017, 2018, 2019	385/991 (38.8%)	142/991 (14.3%)	11.5 ± 9.9	39.2 ± 14.2
B. melanopygus	Pyrobombus	2011, 2014, 2015, 2016, 2018	25/140 (17.9%)	8/140 (5.7%)	27.4 ± 23.9	45.0 ± 13.4
B. mixtus	Pyrobombus	2012, 2015, 2016, 2017	5/14 (35.7%)	1/14 (7.1%)	12.0 ± 10.2	36.0 ± 0.0
B. nevadensis	Bombias	2011, 2013, 2015, 2017	2/20 (10%)	1/20 (5%)	21.5 ± 10.6	56.0 ± 0.0
B. occidentalis	Bombus	2009, 2010, 2011, 2012, 2014, 2015, 2016, 2017, 2018, 2019	119/201 (59.2%)	70/201 (34.8%)	13.7 ± 9.2	35.0 ± 7.1
B. rufocinctus	Cullumanobombus	2011, 2013, 2014, 2015, 2016, 2017, 2018, 2019	3/58 (5.2%)	2/58 (3.4%)	27.7 ± 19.9	38.0 ± 26.9
B. vancouverensis	Pyrobombus	2010, 2011, 2012, 2013, 2014, 2015, 2016, 2017, 2018, 2019	284/761 (37.3%)	86/761 (11.3%)	13.6 ± 9.8	40.2 ± 14.0
B. vandykei	Pyrobombus	2015, 2016	6/31 (19.4%)	4/31 (12.9%)	25.3 ± 15.9	47.0 ± 28.0
B. vosnesenskii	Pyrobombus	2014, 2015, 2016, 2017, 2018	373/774 (48.2%)	195/774 (25.2%)	12.9 ± 9.3	45.5 ± 14.9

Fig. 1.

Timeline of colony development for western North American Bombus species. The gyne collection/installation day indicates the timeline’s starting point, which is represented by the aerial net. Average number of days to first observed brood (nest initiation) from nest installment for each Bombus species indicates the second point in the timeline, which is represented by the brood cells. Average number of days to first observed worker (nest establishment) from nest installment for each Bombus species indicates the final point in the timeline, represented by the bumble bee. The timeline indicates overall colony averages for days to nest initiation and nest establishment. Therefore, time between nest initiation and nest establishment is not equal to individual larval development.

Colonies reared from 2 gynes had significantly higher nest initiation (z-score = 5.59, df = 1, P < 0.001) rates per nest box compared with those reared from a single gyne. However, nest establishment rates did not differ between colonies reared from a single gyne or via cofounding (z-score = 1.05, df = 1, p = 0.15). Colonies reared from a single gyne had a nest initiation rate of 39.5% and a nest establishment rate of 18.5%. Meanwhile, colonies reared via cofounding had a nest initiation rate of 54.1% and a nest establishment rate of 20.8%.

Bombus Colony Size

Colony size parameters varied among species. Worker/drone brood cell production differed significantly among species (F = 28.79, df = 4, P < 0.001) (Fig. 2). Based on pairwise comparisons, B. huntii colonies were significantly larger than B. griseocollis (P < 0.001), B. occidentalis (P < 0.001), and B. vancouverensis Cresson (P < 0.001) colonies. Additionally, B. vosnesenskii colonies were significantly larger than B. griseocollis (P < 0.001), B. occidentalis (P < 0.001), and B. vancouverensis (P < 0.001) colonies. Gyne cell production also differed significantly among species (F = 3.358, df = 4, P = 0.011) (Fig. 3). Based on pairwise comparisons, B. huntii colonies produced significantly more gynes than B. vosnesenskii (P = 0.009).

Fig. 2.

Distribution of worker/drone brood cells produced by Bombus griseocollis, B. huntii, B. occidentalis, B. vancouverensis, and B. vosnesenskii colonies from 2015 to 2018. Crossbars represent the mean.

Fig. 3.

Distribution of gyne brood cells produced by Bombus griseocollis, B. huntii, B. occidentalis, B. vancouverensis, and B. vosnesenskii colonies from 2015 to 2018. Crossbars represent the mean.

Discussion

This observational study provided a comprehensive documentation of the variation in rearing success rates, growth rates, and colony sizes among 15 western North American Bombus species, many of which have not been intensively studied. Furthermore, this study was the first to document rearing of 4 species in captivity, including B. californicus, B. flavifrons Eversmann, B. mixtus Cresson, and B. vandykei Frison, which adds to the literature available on nest success rates and colony development timelines. Overall, these results can be used to maximize captive-rearing success rates and enhance knowledge of systematic nesting biology and colony development in a controlled laboratory environment (Macfarlane et al. 1994, Kwon et al. 2006, Yoneda 2008, Strange 2010, Sarro et al. 2021).

Similar to previous studies, colonies reared via cofounding had significantly higher nest initiation rates per nest box compared to those reared from a single gyne, which supports that pairing gynes increase the probability that one individual will produce brood (Sladen 1912, Plowright and Jay 1966, Bernasconi and Keller 1996, Ptáček et al. 2000, Strange 2010). However, initiation and establishment rates with 2 gynes were not tested systematically across all species. Furthermore, high variability among species was observed in response to this method, which is consistent with Strange (2010). While cofounding can reduce labor, space, and resources needed for initiating colonies, this method should be tested for utility and efficacy in the target species prior to wide-scale implementation, particularly with species of conservation interest.

Across all species, B. rufocinctus and B. nevadensis had the lowest rates of nest success. Furthermore, several species (B. californicus, B. fervidus Fabricius, and B. flavifrons) did not establish a nest (rear a single worker to adulthood); however, this data may be biased due to low sample sizes. Meanwhile, B. griseocollis had the highest rate of nest success (quantified by nest initiation and establishment rates) across the 11-yr period, followed by B. occidentalis, B. vosnesenskii, and B. huntii. Bombus griseocollis initiated and established nesting at a rate of 76.1% and 54.6%, respectively, which was similar to findings from a study by Christman et al. (2022) (initiation: 70.6%; establishment: 52.8%). In addition to having the highest rate of nest success, B. griseocollis also had the fastest nest initiation rate among species, with an average of 8.4 ± 6.7 days to first brood. These results were also consistent with previous research from Christman et al. (2022), which documented an average of 7.6 ± 7 days to first brood for B. griseocollis. We were unable to compare nest establishment between our study and that of Christman et al. (2022) due to differences in the way in which days to nest establishment was calculated. Nest success or developmental timelines for B. occidentalis, B. vosnesenskii, and B. huntii have not been previously documented in the literature, although these species have been commercially produced. This information is likely not available given the proprietary nature of industry produced data.

Similar to our nest initiation and establishment results, colony size parameters varied among the 5 species. Bombus huntii and B. vosnesenskii had the largest colonies, producing an average of 245.5 ± 188.3 and 238.8 ± 151.6 worker/drone brood cells per colony, respectively. Bombus huntii also produced the most gynes, with an average of 13.3 ± 18.4 gyne cells per colony. These findings are of interest given that both of these species are now commercially available in North America, along with B. impatiens Cresson. The fact that they are within the subgenus Pyrobombus, have high rates of nest success, large colony sizes, and, in the case of B. huntii, sizeable gyne production may make them amenable for commercial production. Pyrobombus are pollen storers and have been reported to have lower rates of Vairimorpha bombi infection in wild populations (Sladen 1912; Velthuis and van Doorn 2006, Cordes et al. 2012, Malfi and Roulston 2014, Tripodi et al. 2014). Certain species of Pyrobombus have also been reported to emerge and establish nests early in the season, suggesting that these species have a longer time to persist and produce larger colonies compared with later emerging species (Hobbs 1967). Therefore, early nest establishment may increase colony nest success, making Pyrobombus species more amenable to mass production.

Given that these data were collected opportunistically as colonies were reared in a laboratory for experimental purposes, several biases are present. First, the 15 species were not reared consistently over the 11-yr period and varied in sample size. Over the 11-yr period in which these data were collected, minor changes in rearing protocols occurred to improve success rates, including alterations in bee diet preparation/composition. It is also important to note that the laboratory personnel responsible for rearing the colonies shifted over time. Additionally, given that these colonies were checked at least every 3 days, there is some degree of error associated with the documentation of the first observed brood and first observed worker. However, we assessed the interaction of year and species on nest initiation and establishment rates using generalized linear models and found that there was not a significant time or species effect on success rates. Additionally, the laboratory-reared colonies were used for a variety of studies. Colonies were often selected to be used in these studies based on their comparatively fast nest initiation/establishment rates and large colony sizes within species. Therefore, the reported results for the colony size parameters are likely underestimated. As such, results from this study are purely observational.

While these results are observational, this does not negate the importance of the findings reported in this study. On the contrary, these results enhance knowledge of systematic nesting biology and colony development in a controlled laboratory environment. This can provide insight into wild Bombus nesting biology and colony size, which can be difficult to study given that a majority of their life cycle is spent below ground. Furthermore, these rearing protocols could be employed on a variety of other bumble bees, including imperiled species, to help assist in their recovery. Rearing of ex situ colonies for assisted reintroductions has been identified as a last-resort conservation propagation method to mitigate the effects of population declines, particularly throughout the geographic range in which the species has been extirpated (López-Pujol et al. 2006, Shorthouse et al. 2012, Fritz and Chiari 2013, IUCN/SSC 2013; Draper et al. 2020). Therefore, captive rearing of bumble bees could prove to be a powerful tool in bumble bee conservation efforts.

Author Contributions

James Strange (Conceptualization-Lead, Data curation-Equal, Funding acquisition-Lead, Investigation-Equal, Methodology-Lead, Project administration-Lead, Resources-Lead, Supervision-Lead, Writing – original draft-Equal), Amber Tripodi (Data curation-Lead, Investigation-Equal, Methodology-Equal, Project administration-Lead, Supervision-Lead, Writing – review & editing-Equal), Thuy-Tien Lindsay (Data curation-Equal, Investigation-Equal, Methodology-Equal, Writing – review & editing-Equal), James D. Herndon (Data curation-Equal, Investigation-Equal, Methodology-Equal, Validation-Equal, Writing – review & editing-Equal), Joyce Knoblett (Data curation-Equal, Investigation-Equal, Methodology-Equal, Resources-Equal), Morgan Christman (Formal analysis-Equal, Investigation-Equal, Software-Equal, Validation-Lead, Visualization-Equal, Writing – original draft-Equal), Nezhat Barkan (Formal analysis-Equal, Software-Equal, Visualization-Equal, Writing – original draft-Equal), Jonathan Koch (Investigation-Equal, Writing – review & editing-Equal)

Conflict of Interest

The authors declare no conflicts of interest. The authors have no known competing financial interests or relationships with individuals or organizations that could affect or bias scientific judgment or the ability to publish, analyze, discuss, or interpret the data reported in this research.

Data Availability

The data and code supporting the findings of this study are openly available on Zenodo at https://doi.org/10.5281/zenodo.7596923.