ArticlePDF Available

Honey bee drones are synchronously hyperactive inside the nest

July 2023
Animal Behaviour 203(6)

DOI:10.1016/j.anbehav.2023.05.018

Authors:

Jacob Davidson

Max Planck Institute of Animal Behavior

Benjamin Wild

Freie Universität Berlin

David Michael Dormagen

Freie Universität Berlin

Show all 7 authorsHide

Spatial location of drones and workers. Nest contents over the observation period and two-dimensional spatial histograms showing the locations of workers and drones each day, at 5-day intervals (see Appendix, Fig. A1 for all days). For each image, we display position in the observation hive by showing the back side on the left and the front side on the right; in this image, the nest exit is on the lower right corner (see also Appendix, Fig. A2).

…

Figure A8. Daily weather for observation days. Nearby and regional weather data for observation days, showing average temperature, sunshine duration, total precipitation, average atmospheric pressure, average wind speed and average vapor pressure deficit for each day. Not all measurements were available for all weather stations. Recorded values of precipitation and wind speed sometimes differed for the different weather stations, but other measurements were similar. Nearby station Konstanz-Litzelstetten data are from MeteoBlue (2.2 km from hive), and other data are open data from the Deutsche Wetterdienst (DWD). Station Konstanz was the closest at 1.6 km from the hive.

…

Figures - uploaded by Jacob Davidson

Content may be subject to copyright.

Content uploaded by Jacob Davidson

Content may be subject to copyright.

Honey bee drones are synchronously hyperactive inside the nest

Louisa C. Neubauer

, Jacob D. Davidson

, Benjamin Wild

David M. Dormagen

, Tim Landgraf

, Iain D. Couzin

, Michael L. Smith

Department of Collective Behaviour, Max Planck Institute of Animal Behavior, Konstanz, Germany

Department of Biology, University of Konstanz, Konstanz, Germany

Centre for the Advanced Study of Collective Behaviour, University of Konstanz, Konstanz, Germany

Department of Mathematics and Computer Science, Freie Universit€

at Berlin, Berlin, Germany

Department of Biological Sciences, Auburn University, Auburn, AL, U.S.A.

article info

Article history:

Received 15 October 2022

Initial acceptance 2 December 2022

Final acceptance 1 May 2023

Available online xxx

MS. number: A22-00492R

Keywords:

automated tracking

drone

in-nest behaviour

reproductive behaviour

superorganism

Eusocial insects operate as an integrated collective with tasks allocated among individuals. This applies

also to reproduction, through coordinated mating ﬂights between male and female reproductives. While

in some species male sexuals take only a single mating ﬂight and never return, in the western honey bee,

Apis mellifera, the male sexuals (drones) live in the colony throughout their lives. Prior research has

focused almost exclusively on drone behaviour outside the nest (mating ﬂights), while ignoring the

majority of their life, which is spent inside the nest. To understand the in-nest behaviour of drones across

their lives, we used the BeesBook tracking system to track 192 individually marked drones continuously

for over 20 days, to examine how drones moved and spent time in the nest. In agreement with previous

work, we found that drones spend most of their time immobile at the nest periphery. However, we also

observed that drones have periods of in-nest hyperactivity, during which they become the most active

individuals in the entire colony. This in-nest hyperactivity develops in drones after age 7 days, occurs

daily in the afternoon and coincides with drones taking outside trips. We found strong synchronization

across the drones in the start/end of activity, such that the drones in the colony exhibited a ‘shared

activation period’. The duration of the shared activation period depended on the weather; when con-

ditions were suitable for mating ﬂights, the activation period was extended. At the individual level,

activation order changed from day to day, suggesting that both the external inﬂuence of weather con-

ditions as well as exchange of social information inﬂuenced individual activation. Using an

accumulation-to-threshold model of drone activation, we show that simulations using social information

match experimental observations. These results provide new insight into the in-nest behaviour of drones

and how their behaviour reﬂects their role as the male gametes of the colony.

The parts of an organism can be deﬁned by their functional role

(e.g. gas-exchange, digestion, locomotion) or by their ultimate role:

soma versus gametes (Weismann, 1893). This same concept applies

to eusocial insect colonies; workers can be organized into func-

tional subgroups that carry out tasks, similar to organs, but colonies

can also be deﬁned according to those that form the soma

(workers) versus gametes (sexuals) (Beshers &Fewell, 2001;

H€

olldobler &Wilson, 2009;O'Shea-Wheller et al., 2021;Seeley,

1989;Smith &Szathm

ary, 1995). Male reproductives in honey

bee colonies, drones, are the equivalent of colony level sperm,

whose goal is to mate with virgin queens from other colonies

(Koeniger et al., 2005;Loper et al., 1992;Ruttner &Ruttner, 1966;

Woodgate et al., 2021). Drones are important for a colony's repro-

ductive success, and each virgin queen mates with multiple drones

to obtain sperm from diverse patrilines, a critical contribution for

colony level function (Jones et al., 2004;Mattila et al., 2012;Mattila

&Seeley, 2007;Tarpy, 2003).

In many social insect systems, sexuals are reared and depart

only on a single mating ﬂight, never to return (e.g. Solenopsis

invicta:Tschinkel, 2006;Reticulitermes speratus and Coptotermes

formosanus:Mizumoto &Dobata, 2019). Across bees, male mating

strategy depends on patterns of female emergence, nest distri-

bution and male density (Paxton, 2005). Mating strategies deﬁne

male behaviour, but most studies focus on behavioural outcomes

outside the nest, while neglecting their in-nest behaviour. In the

honey bees, genus Apis, drones live their entire lives in/on the

*Corresponding author.

E-mail address: mls0154@auburn.edu (M. L. Smith).

These authors contributed equally to this work.

Contents lists available at ScienceDirect

Animal Behaviour

journal homepage: www.elsevier.com/locate/anbehav

https://doi.org/10.1016/j.anbehav.2023.05.018

Animal Behaviour xxx (xxxx) xxx

Please cite this article in press as: Neubauer, L. C., et al., Honey bee drones are synchronously hyperactive inside the nest, Animal Behaviour

(2023), https://doi.org/10.1016/j.anbehav.2023.05.018

nest, departing on multiple mating ﬂights per day (Oldroyd &

Wongsi ri , 20 06;Seeley, 2019). Therefore, even if their ultimate

purpose is solely related to colony reproduction, the fact that they

depart and repeatedly return shows that drones are part of the

colony's overall organization within the nest. Previous research

has examined both the in-nest and out-of-nest behaviour of

workers, who perform the bulk of colony functions (H€

olldobler &

Wilson, 2009;Seeley, 1995,2010), but prior work on drones has

focused almost exclusively on their behaviour outside the nest,

speciﬁcally their mating ﬂights.

Drone mating ﬂights occur across all species of Apis (Oldroyd &

Wongsiri, 2006). Each Apis species has its own time window for

drone departures, which helps reproductive isolation in regions

where multiple species of Apis coincide (Koeniger et al., 1988;

Oldroyd &Wongsiri, 2006;Wongsiri et al., 1997). In the western

honey bee, A. mellifera, drones depart on multiple mating ﬂights

each afternoon, during which they visit multiple drone congrega-

tion areas (Oertel, 1956;Reyes et al., 2019;Ruttner, 1966;Taber,

196 4;Woodgate et al., 2021). Drones that successfully mate will

die, but there are thousands of drones produced for each available

virgin queen (Smith et al., 2016), so most drones depart on daily

ﬂights without ever mating. Flight activity depends on the drone's

maturation; young drones perform short orientation ﬂights (Oertel,

1956;Witherell, 1971), and older drones depart on long-distance

mating ﬂights 1e5 km from the nest (Reyes et al., 2019;Ruttner

&Ruttner, 1972). Previous work has shown that drone ﬂights

tend to occur only under speciﬁc weather conditions: temperature

above 20



C, light intensity around 2000 lx, no cloud cover and

wind speed below 30 km/h (Koeniger, 1986;Neves et al., 2011;

Reyes et al., 2019). These environmental conditions align with the

conditions needed for virgin queens to depart on mating ﬂights

(Koeniger, 1986;Taber, 1964). Indeed, virgin queens need to be

particularly choosy about weather conditions; if they fail to return

home, the colony is rendered ‘hopelessly queenless’and will perish

(Smith, 2018). At the end of the summer, when there are no more

virgin queens with which to mate, drones are no longer useful to

the colony and workers evict them to die of starvation outside the

nest (Cicciarelli, 2013;Free &Williams, 1975;Morse et al., 1967;

Wharton et al., 2008).

These details describe the outside activities of drones, but less

is known about their behaviour within the nest. Previous work

found that drones tend to remain stationary on the combs, only

moving to feed (Free,1957). Young drones tend to be located in the

centre brood-rearing region of the nest, whereas older drones are

located at the nest periphery (Fukuda &Ohtani, 1977). Drones do

not perform colony work, although they contribute to nest ther-

moregulation by heating themselves when cold (Kovac et al.,

2009).

Here, we examine the in-nest behaviour of drones of the

western honey bee, by tracking the movement patterns and spatial

positioning of individual drones in an observation hive over their

entire lives. We found that drones exhibited a daily period of in-

nest hyperactivity, during which they also took trips outside of

the nest. The duration and onset of the high-activity period

depended on weather conditions. Looking at individual drones, we

found that the start and end of the activation period was highly

synchronized and that individual drone activation order was not

consistent from day to day. Finally, we formulated an

accumulation-to-threshold model of drone activation and exam-

ined simulation results with and without social information, in

comparison to experimental observations. These results provide

new insight into the in-nest behaviour of drones as the male

reproductive units of the colony.

METHODS

Observation Hive and Recording

This study was performed at the University of Konstanz, Ger-

many (47



N, 9



E). On 10 May 2019, the observation

hive was installed with a single queen, 4000 unmarked workers

and three frames containing brood and honey (observation hive

dimensions: 490 742 mm; ‘Deutsche-Normal’frames:

395 225 mm). From 5 June to 23 September 2019, individually

marked newborn workers were introduced to the observation hive

colony every 4e5 days (250e400 newborns per introduction). On 7

June and 12 June, when colonies were producing drones, we indi-

vidually marked and introduced drones to the same observation

hive colony (192 total: 160 and 32, respectively). Newborn bees

were sourced from colonies headed by naturally mated queens

from the University of Konstanz apiary (Apis mellifera carnica). All

newborns were hatched overnight in an incubator, so the size of the

cohort matched the number of individuals that would have natu-

rally emerged overnight. The incubator was set to 34



C and 50%

relative humidity (RH), and newborns were marked that morning

with individual BeesBook tags (Boenisch et al., 2018;Wario et al.,

2015). BeesBook tags are printed on paper and attached to the

bee's thorax, where they remain attached for the life of the bee.

From 5 June to 23 October 2019, the observation hive was

recorded at three frames/s with four Basler acA4112-20um cameras

ﬁtted with Kowa LM25XC lenses. To mimic natural conditions in-

side a nest, the colony was illuminated with infrared light (850 nm,

3 W LEDs), which the bees cannot perceive (Peitsch et al., 1992),

and the colony had free access to the outside via a 2 cm diameter

entrance tunnel. This investigation focuses on the recording period

during which drones were present in the colony, which began 7

June 2019.

To create a map of the nest, we periodically traced the contents

of the observation hive onto plastic sheets by outlining the

following: honey storage, pollen storage, brood, empty comb,

wooden frames, peripheral galleries and dances observed on the

dance ﬂoor (as in Smith et al., 2016). These plastic sheets were then

scanned with an architectural scanner (Ruch-Medien, Konstanz)

and digitized.

Ethical Note

The experimental procedures were designed to minimize po-

tential impact on the animal subjects. Workers and drones were

tagged using paper tags, which weighed less than 1% of the total

weight of the bee, and a bee-safe adhesive. Tagged individuals were

introduced to the colony to minimize stress: they were fed ad

libitum, had multiple hours of acclimatization and were allowed

self-directed entry into the colony. These experiments were

inherently noninvasive and involved no manipulations that would

incur additional stress upon individuals or the colony.

Data Processing

Using the BeesBook tracking system, the raw images were

processed to detect and decode individually marked bees (Wario

et al., 2015;Wild et al., 2021). These processed data contain tag

identity (ID), ID detection conﬁdence, position in the nest, bee

orientation and time. All data were stored in a PostreSQL database.

The death date of each marked individual was estimated using a

Bayesian change-point model (as in Wild et al., 2021). This

method accounts for a low rate of erroneous detections in bees

L. C. Neubauer et al. / Animal Behaviour xxx (xxxx) xxx2

Please cite this article in press as: Neubauer, L. C., et al., Honey bee drones are synchronously hyperactive inside the nest, Animal Behaviour

(2023), https://doi.org/10.1016/j.anbehav.2023.05.018

that have already died and time periods when individuals are

observed less frequently or not at all (e.g. while foraging). An in-

dividual's death date was used as a cutoff for including data in

subsequent calculations. Given that drones have a low probability

of mating success per ﬂight (thousands of drones are produced

per virgin queen; Smith et al., 2016) and a high potential for

mortality (Visscher &Dukas, 1997), we did not attempt to differ-

entiate drones that successfully mated and died versus those that

simply died while outside the nest or within the nest of other

causes.

We used the processed trajectory data to calculate metrics

representing the behaviour of each individual. We did this by

averaging over time bins of duration T,whereT¼5 min was used

for all analyses shown here. All data points used in the analysis

were above a detection conﬁdence threshold of 0.8 (detection

conﬁdence is an output of the BeesBook tracking system that in-

dicates how likely a detection is correct; see Boenisch et al., 2018).

We calculated behavioural metrics for each bee that had a mini-

mum of 10 detections in a time bin. In general, we made this

choice in order to keep as much data as possible during the data

processing step, while also ﬁltering out likely errors and main-

taining ﬂexibility in possible analyses with the data. For the

analysis of drone behaviour, we focus on speed and time spent

outside; however, the full list of all calculated metrics included

quantities to represent space use within the nest (time on honey,

time on brood, time on dance ﬂoor and distance from exit),

detection (time observed, time outside, number of outside trips,

number of dance ﬂoor visits) and movement/spatial localization

(speed, circadian coefﬁcient, dispersion, fraction of nest visited;

see Smith et al., 2022 for how these metrics are calculated). For

each day, we calculated behavioural parameters for each bee,

averaged in the bins of dura tion T(7 June and 12 June cohorts: 736

workers and 192 drones, total). For all averages shown in the

ﬁgures, we used a per-bee average of the calculated metrics.

To estimate when a bee was outside, we used an algorithm with

input from the binned calculations for detection and exit distance.

Because we did not have direct observations or detections corre-

sponding to exiting the hive (i.e. there was no camera at the hive

exit to read barcodes), leaving the nest meant that there would be

gaps in the detection of individual bees for some period. However,

it is also possible that a bee's barcode in the observation hive was

not always detected, for example, if the bee was upside down or in a

dense crowd of other bees. Therefore, we used both detection and

exit distance to estimate when bees were outside the nest. We ﬁrst

calculated the time observed and the median exit distance in 5 min

bins over the course of a day. A bee was then estimated to have

exited the nest in a time bin t

exit

if the time observed in t

exit

was less

than a threshold of t

obs

¼2 s and if the median exit distance in time

bin t

exit

1 was less than a threshold d

exit

¼31.25 cm (2500 pixels).

The bee was considered to have re-entered in bin t

enter

if the time

observed in t

enter

was greater than or equal to t

obs

. The values t

obs

and d

exit

are analysis parameters, and while the results and average

trends are not sensitive to the values used, the quantitative esti-

mates of outside time can depend strongly on the choice of d

exit

.We

chose the value of 31.25 cm to represent a feasible median exit

distance for a bee travelling to the exit during a 5 min period (see

Appendix, Fig. A2 for exit distances labelled within the observation

hive). With these results, we determined multiple instances of exit

and re-entry times during the course of a day and used this to

calculate the number of outside trips (the number of times a bee

was estimated to have exited the nest), as well as the time spent

outside.

RESULTS

In-hive Movement of Drones Over Time

We individually tagged and introduced 192 newly emerged

drones into an observation hive and tracked their motion contin-

uously at three frames/s over 20þdays using the BeesBook auto-

mated tracking system (Boenisch et al., 2018;Wario et al., 2015;

Wild et al., 2021). The drones were divided into two cohorts with

birth dates 5 days apart. Two worker cohorts were introduced

simultaneously with the drones (bees within each cohort had the

same date of birth; worker/drone cohort 1: 7 June; worker/drone

cohort 2: 12 June; mid-June is when colonies naturally produce

drones). We quantiﬁed the individual behaviour of drones by

calculating the median speed of individuals in 5 min bins, as well as

using position and detection within the observation hive to esti-

mate when individuals left the nest (see Methods). We used nest

tracing to map the comb contents in the observation hive (honey

stores, brood care, pollen stores, festoon for comb building; as in

Smith et al., 2016,2022), so that the position of bees in the hive

could be compared with current nest structure and use.

In agreement with previous observations (Free, 1957), we found

that drones remained immobile in the nest throughout most of the

day (per-drone median speed: 0.08 ±0.027 cm/s; per-worker

speed for workers in cohorts 1 and 2: 0.262 ±0.035 cm/s; two-

sample ttest: P<0.001). However, we also found that drones had

extreme bouts of hyperactivity (Fig. 1a). During these hyperactive

bouts, drone speed was several times faster than worker speed

(per-drone median speed during shared activation periods:

0.779 ±0.282 cm/s; per-worker speed for workers in cohorts 1 and

2 during shared activation periods: 0.355 ±0.068; two-sample t

test: P<0.001; see below and Methods for deﬁnition of shared

activation period). This demonstrates that while drones did spend

the majority of their time immobile, or barely moving, they could

also be the fastest individuals in the colony, albeit during a limited

time window (see trajectories in Supplementary Video S1).

The increase in drone speed was ﬁrst detectable at age 5e6 days,

became pronounced by age 7 days and continued throughout the

drone's life (Fig. 1a). The timing of the hyperactive bouts corre-

sponded to when drones moved near to the nest entrance (Fig. 1c)

and spent time outside the hive (Fig. 1b), and so is likely associated

with the onset of orientation/mating ﬂights. These bouts tended to

occur in afternoons, when outside temperatures were warm and

the sun was out (Fig. 1d).

Drones also changed their position in the nest as they aged.

Young drones (ca. <10 days old) were located in the centre of the

nest, on comb containing brood (Fig. 2,Appendix, Fig. A1). From ca.

age 10 days, drones shifted from the centre of the nest to the pe-

riphery, i.e. the edge areas of each frame in the observation hive

(Appendix, Fig. A3), which corresponded to areas bordering the

brood. As they aged further (age 14e19 days), drones shifted closer

to the nest entrance (lowest frame; see Fig. 2); however, their

average distance from the exit changed during the day (Fig. 1c).

After drone cohort 2 had reached age 10 days (22 June), the two

drone cohorts used similar areas of the nest. The shift from upper/

periphery areas to the lower frame was noticeable for both cohorts

on 26 June (age 19 and 14 days for drone cohorts 1 and 2, respec-

tively; see Appendix, Fig. A1). This suggests that spatial positioning

of drones in the nest is inﬂuenced not only by age, but also by the

positioning of other drones.

During the observation period, the colony was also in the pro-

cess of forming a festoon to build comb on the upper frame (see red

L. C. Neubauer et al. / Animal Behaviour xxx (xxxx) xxx 3

Please cite this article in press as: Neubauer, L. C., et al., Honey bee drones are synchronously hyperactive inside the nest, Animal Behaviour

(2023), https://doi.org/10.1016/j.anbehav.2023.05.018

‘Festoon’areas in Fig. 2, and in the Appendix, Fig. A1). The last day

the festoon was observed was 26 June. After this, there was a

notable increase in honey stores in the nest; honey areas went from

10.1% of the nest on 26 June to 12.1% on 28 June, and then to 17.2%

on 2 July. During this time of increasing honey stores, all worker

bees increased their time spent outside (the fraction of all workers

outside the hive after 26 June; Fig. 1b).

Synchronized Activation of Drones

We observed synchronization in drone activity, such that many

drones became active at the same time. We refer to this as a ‘shared

activation period’(s

act

). We used the data to form a working deﬁ-

nition of this period to enable quantitative analysis. To do this, we

denoted an individual drone as ‘activated’in time bin tif its median

speed was greater than a threshold of s

act

, or if the drone was

identiﬁed as being outside the nest during t. We used a high value

of median speed for the threshold: s

act

¼0.5 cm/s, which corre-

sponds to the 92.8% quantile of median speed values for all tracked

bees in the observation period of 7e29 June. We deﬁned activation

as a bee having either high speed or being outside of the nest

because these activities coincided and they are mutually exclusive

(i.e. in-nest speed does not exist when drones are outside the nest).

To compare the onset and duration of activity on different days, we

then deﬁned the ‘shared activation period’as when >25% of drones

age 7 days were activated at the same time (Fig. 3a). We used the

age threshold of 7 days old because drone activation was not

strongly observed until this age. Note that we deﬁned the shared

activation period to enable quantitative analysis; we chose pa-

rameters to reﬂect a meaningful representation of the data during

this period (Fig. 3a) and our deﬁnition was not overly sensitive to

the exact parameter values for speed threshold or fraction of drones

(Appendix, Fig. A4).

After a shared activation period was ﬁrst observed on 14 June, all

subsequent days (14e29 June) had a shared activation period with

the exception of 20 June. Although the shared activation period

varied from day to day, it always began between 1300 and 1600

hours and always ended by 1800 hours. On some days, as many as

70% of the drones were active during this period. The duration of

the shared activation period varied from 3 h (23 June) to only

15 min (21 June) (Fig. 3).

Individuals in drone cohort 2 started from age 7 days (19 June)

to increase their speed and take part in the shared activation

period, becoming activated at the same time as drone cohort 1 each

day (Fig. 3b, c). Therefore, even though these different drone co-

horts were of different ages, they were synchronized in activity

such that we observed only a single shared activation period.

During the same time periods, worker bees did not exhibit a

particular increase in activity levels (Appendix, Figs A6eA7).

The duration and onset of the shared activation period depen-

ded on daily weather conditions. Of the quantities examined,

sunshine duration had the highest positive correlation with dura-

tion of the shared activation period (Fig. 4). On one day (20 June),

we did not observe a shared activation period; the drones remained

0.2

0.4

0.6

Fraction out

of hive

(b) Out of hive

(d)

78 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

Day of June 2019

Exit distance

(cm)

Temp.

(ºC)

30 Temperature

of June 2019

Intensity

(W/m2)

78910 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

500

Sunshine intensity

Drones, cohort 2

Drones, cohort 1

Workers, cohort 2

Workers, cohort 1 All worker bees

1,1

Age: Age:

0.5

1.5

Median speed

(cm/s)

(a) Median speed

11,

Age:

11,

16,

Age:

16,

21,

Age:

(days)

Figure 1. Movement of drones and workers, and weather over time: (a) median speed, (b) fraction of bees out of the hive and (c) median exit distance for each drone and worker

cohort over the observation period. Drone and worker cohorts 1 were introduced on 7 June; drone and worker cohorts 2 were introduced on 12 June. Each line shows the per-bee

average across the group, calculated using time bins of 5 min (see Methods for data processing details). The black line shows values for all marked worker bees in the observation

hive (this includes 3246 bees during this time period). (d) Weather data of temperature and sunlight intensity for local weather station MeteoBlue-Litzelstetten (2.2 km from hive),

plotted over time using the same time range as (a) e(c).

L. C. Neubauer et al. / Animal Behaviour xxx (xxxx) xxx4

Please cite this article in press as: Neubauer, L. C., et al., Honey bee drones are synchronously hyperactive inside the nest, Animal Behaviour

(2023), https://doi.org/10.1016/j.anbehav.2023.05.018

immobile throughout the day. Using local weather data, this day

had fewer hours of sunlight (<5 h of sunlight recorded at both

nearby weather stations: DWD-Konstanz (1.6 km from hive) and

MeteoBlue-Litzelstetten (2.2 km from hive); see Appendix, Fig. A8).

On the following day (21 June), some drones did increase their

speed and took trips outside (over 40% of drones became activated

for a brief period on this day; see Fig. 3a), but this day had the

shortest activation period (15 min). On each day, the onset of the

shared activation period occurred slightly after the period of

strongest sunlight (Appendix, Fig. A9).

Individual Drone Activity Onset

We next examined the activity patterns of individual drones. We

saw remarkable synchronization in the onset and end of hyperac-

tivity and outside trips among the individual drones (Fig. 5). During

active periods, drones exited the nest multiple times and main-

tained high movement speed when they returned to the nest. The

number of outside trips per drone during the shared activation

period was 1.17 ±1.05, meaning that most drones took one to two

trips (although some drones did exit many more times; see

Appendix, Fig. A5a). The average time per outside trip during the

activation period was 24.2 ±11.6 min. Drones tended to increase

their speed before exiting the nest (Appendix, Fig. A5b). The end of

the activation period saw individuals returning into the nest and

reducing their speed, showing that not only the onset of activity,

but also the decrease, was synchronized among drones (Fig. 5).

To determine whether the drones were consistent in their onset

of activity (e.g. drones that were consistently ‘early-to-activate’or

‘late-to-depart’), we deﬁned the individual activation time for

drone ias the ﬁrst time bin tstarting from 1 h preceding the shared

activation period where the drone's speed s

act

or the drone was

outside the hive. We then calculated the correlation in activation

timing of all drones from one day to the next. The correlation was

nonzero and varied across days but was generally low (Fig. 6a). This

suggests that certain drones did not show a strong tendency to

consistently be ﬁrst or last to become activated or to consistently

‘lead’the onset of shared activation each day.

We furtherasked whether previous behaviour or location predicts

activation timing of individuals with respect to the shared activation

period. We used the morning hours of 0000e1200 hours each day to

calculate average values of exit distance and speed for each drone.

Because drones changed their location in the nest over time (see

Fig. 2), we used a ranking of drones within each day. Figure 6bshows

that there was nearly no correlation of an individual's activation

timing with either morning period exit distance or speed.

Model of Social Activation of Drones

The synchronization of the activation period (Fig. 5)suggested

that social information exchange, in addition to external weather-

related factors, drives individual activation. To consider this mech-

anism in detail, we formulated a simple threshold model and

simulated the response of a group of agents with and without social

8 June 13 June 18 June 23 June 28 June

20 cm

Age 1

Drones, cohort 1

Drones, cohort 2

Workers, cohort 1

Workers, cohort 2

Age 6 Age 11 Age 16 Age 21

Age 1 Age 6 Age 11 Age 16

Age 21

Honey

Capped brood

Young brood

Empty comb

Pollen stores

Wooden frames

Festoon

Blank space

Figure 2. Spatial location of drones and workers. Nest contents over the observation period and two-dimensional spatial histograms showing the locations of workers and drones

each day, at 5-day intervals (see Appendix, Fig. A1 for all days). For each image, we display position in the observation hiveby showing the back side on the left and the front side on

the right; in this image, the nest exit is on the lower right corner (see also Appendix, Fig. A2).

L. C. Neubauer et al. / Animal Behaviour xxx (xxxx) xxx 5

Please cite this article in press as: Neubauer, L. C., et al., Honey bee drones are synchronously hyperactive inside the nest, Animal Behaviour

(2023), https://doi.org/10.1016/j.anbehav.2023.05.018

information exchange. The purpose of this model was not to spe-

ciﬁcally match the observed results, but rather to determine what

conditions would be necessary to observe the same general behav-

ioural patterns (e.g. presence/absence of social information,

individual thresholds, noise levels). For other examples of threshold

models applied to social insects, see Beshers and Fewell (2001),Jandt

and Dornhaus (2014) and Ulrich et al. (2021). In the model, an in-

dividual agent ibecomes activated when its internal decision state x

Time of day (hours)

Fraction of drones

13 June 14 June 15 June 16 June 17 June 18 June

19 June 20 June 21 June 22 June 23 June 24 June

25 June 26 June 27 June 28 June 29 June Out of hive or speed

over threshold

Out of hive

Speed over threshold

Shared activation

threshold

Shared activation period

(a)

Time of day (hours)

1300

1400

1500

1600

1700

1800

0.2

0.4

0.6

1300

1400

1500

1600

1700

1800

1300

1400

1500

1600

1700

1800

1300

1400

1500

1600

1700

1800

1300

1400

1500

1600

1700

1800

1300

1400

1500

1600

1700

1800

1300

1400

1500

1600

1700

1800

0.2

0.4

0.6

1300

1400

1500

1600

1700

1800

1300

1400

1500

1600

1700

1800

1300

1400

1500

1600

1700

1800

1300

1400

1500

1600

1700

1800

1300

1400

1500

1600

1700

1800

1300

1400

1500

1600

1700

1800

0.2

0.4

0.6

1300

1400

1500

1600

1700

1800

1300

1400

1500

1600

1700

1800

1300

1400

1500

1600

1700

1800

1300

1400

1500

1600

1700

1800

0000

0100

0200

0300

0400

0500

0600

0700

0800

0900

1000

1100

1200

1300

1400

1500

1600

1700

1800

1900

2000

2100

2200

2300

Day of June 2019

Fraction drones out of the hive (0-1)

Median speed (0-2.5 cm/s)

(b) (c)

Shared activation period

Drones, cohort 1

Drones, cohort 2

0000

0100

0200

0300

0400

0500

0600

0700

0800

0900

1000

1100

1200

1300

1400

1500

1600

1700

1800

1900

2000

2100

2200

2300

Drones, cohort 1

Drones, cohort 2

Figure 3. Shared activation period and hourly activity. (a) We deﬁned the ‘shared activation period’(s

act

) to begin when >25% of drones age 7 days had high speed (thresh-

old >0.5 cm/s) or were outside of the hive and to end the last time when <25% of drones satisﬁed either condition. Starting from 14 June, every day had a shared activation period,

with the exception of 20 June. Note that on 25e28 June, we were able to detect the onset of the shared activation period, but the end of the period was obscured by the comb

contents measurements (see also Appendix, Fig. A4 for the sensitivity of the collective activation period to the parameters of the fraction of drones and s

act

). (b) Median speed,

averaged across each cohort (Yaxis range 0e2.5 cm/s for each day), and (c) the fraction of drones outside the hive, for each cohort (Yaxis range 0e1 for each day). Each day is aligned

by hour (Xaxis). Gaps in the data shown are due to the comb contents measuring period, which requires the cameras to be temporarily paused.

L. C. Neubauer et al. / Animal Behaviour xxx (xxxx) xxx6

Please cite this article in press as: Neubauer, L. C., et al., Honey bee drones are synchronously hyperactive inside the nest, Animal Behaviour

(2023), https://doi.org/10.1016/j.anbehav.2023.05.018

reaches a threshold of

.EachoftheNagents in the group can have a

different threshold value. An individual's decision state is described

by an OrnsteineUhlenbeck process (Smith, 2000), as a leaky accu-

mulator of social information and an external signal, plus noise:

¼dtð

iþð1

ÞSðtÞx

Þþ

dWðtÞ:(1)

In this equation,

is the timescale for changes in the decision

variable,

is the amplitude of the noise in the decision variable and

dW(t) is a Wiener noise process. The parameter

sets the weighting

of social information versus the external signal. Social information is

represented by the fraction of group members that have already

crossed their decision thresholds, i.e. Cy

D, where C$Drepresents an

average, and y

¼1ifx



and otherwise 0. The external signal is

set to S(t)¼Asin(2

t) in order to represent periodic changes such as

daily rhythms or sunlight, and the amplitude Asets the strength of

the signal (see Fig. 7a for an illustration of the model, showing the

external signal and decision states of three simulated agents).

We simulated different parameter settings for noise, threshold

differences among group members and social information use to

ask what mechanisms are consistent with the main observations in

the data. The parameter

represents noise in an individual's de-

cision process (‘decision noise’). We set individual agent threshold

values using the distribution

¼N(0.5,

), such that

repre-

sents ‘threshold differences’among individuals. For different

combinations of decision noise and threshold differences, we

17.5 20 22.5 25

Avg. temperature (ºC)

0.5

1.5

2.5

Duration of shared

activation period (h)

13 June

20 June

r=0.41

2.5 5 7.5 10 12.5 15

Sunshine duration (h)

13 June

20 June

r=0.665

2.5 5 7.5 10 12.5

Vapor pressure deficit (hPa)

13 June

20 June

r=0.479

Figure 4. Duration of shared activation period and daily weather conditions. Duration of the shared activation period (Yaxis) is plotted as a function of daily weather averages or

totals (Xaxis). Shown are average daily temperature, total sunshine duration and average vapor pressure deﬁcit obtained from MeteoBlue weather station Konstanz-Litzelstetten

(see also Appendix, Figs A8eA9 for more details and other weather measurements during the observation period, including additionally precipitation, pressure and average wind

speed). The data point for 13 June is labelled to show the date before drones became activated; 20 June is labelled to show the date that did not have a shared activation period. The

correlation values (r) are shown in the upper right of each plot. Using a Pvalue threshold of 0.05, only the correlation with total sunshine duration was signiﬁcant (Pvalues of 0.102,

0.004 and 0.052 for temperature, sunshine duration and vapor pressure deﬁcit, respectively).

Time of da

(hours)

Median speed

Day of June 2019

1200 1300 1400 1500 1600 1700 1800

Speed

(cm/s)

1.75

1.5

1.25

0.75

0.5

0.25

(a) 13

Out of hive

1200 1300 1400 1500 1600 1700 1800

Comb

measurement

period

Shared

activation

start/end

In hive (grey)

Out of hive

(b)

Figure 5. Individual activity patterns. Raster plots, where each row represents an individual drone on a particular day, showing (a) speed and (b) trips out of the hive, in 5 min bins

(see Methods for data processing details). White areas are when cameras were temporarily blocked to transcribe nest contents, and red lines denote the start/end of the shared

activation period (see also Fig. 3). On 25e28 June, we were able to detect the onset of the shared activation period, but the end of the period was obscured by the comb contents

measurements (see Appendix, Fig. A7 for analogous plots with worker cohorts 1 and 2).

L. C. Neubauer et al. / Animal Behaviour xxx (xxxx) xxx 7

Please cite this article in press as: Neubauer, L. C., et al., Honey bee drones are synchronously hyperactive inside the nest, Animal Behaviour

(2023), https://doi.org/10.1016/j.anbehav.2023.05.018

compared simulations of zero versus nonzero social information

exchange (as set by the parameter

In the data, we observed that the shared activation period was

longer on days with more sunshine (Fig. 4). The signal amplitude A

can be considered to represent the strength of the sun during the

day (orange parabola in Fig. 7); for all parameter values, the model

showed the trend of increasing activation time with higher A.We

observed, however, that the synchronization and time of onset of

the activation period depended on the values of noise, threshold

differences and social information weight (Fig. 7b, c). With zero

decision noise and all agents having the same threshold (

¼0,

¼0), we observed perfect synchronization among agents,

regardless of social information use: all group members were either

activated or not activated (top left subplot in Fig. 7b, c). A more

realistic case considers noise in the decision-making process to

represent individuals as imperfect sensors of their environment

(

>0; bottom row of Fig. 7b, c). If either decision noise or

threshold differences are present, then the overall fraction of the

group activated showed gradual changes with signal amplitude, i.e.

no longer a synchronized all-or-none response (Fig. 7b).

The use of social information increased the synchronization of

activation among group members. Even with decision noise and

(b)

0 50 100 150

Rank of median exit distance

avera

ed across 0000-1200

–50

100

Activation time minus shared

activation start time (min)

Rank of median speed

avera

ed across 0000-1200

0 50 100 150

0.1

0.2

0.3

Corr. indiv. activation

timing with previous day

Day of June 2019

(a)

Figure 6. Testing for consistency in activity onset. (a) Correlation of individual activation time from one day to the next. The value for a designated day is the Pearson correlation of

the timing of drones that became activated on that day, with the timing of drones that had also become activated on the previous day. Because there was no shared activation period

observed on 20 June, the correlation was not calculated for 20 and 21 June. (b) Scatter plot showing ranked average morning time (0000e1200 hours) median exit distance (left) or

speed (right) versus individual activation time. All days with shared activation periods are included; each point represents one drone on a single day. The ranked value of exit

distance or speed is with respect to other drones on that day; this was done because drone positioning changes over developmental time (see Fig. 2). The overall correlations were

low: the correlation of ranked exit distance with individual activation time was 0.051 and that of ranked speed with individual activation time was 0.017.

(b)

0.5

Fraction activated

0000

0200

0400

0600

0800

1000

1200

0000

0200

0400

0600

0800

1000

1200

0000

0200

0400

0600

0800

1000

1200

0000

0200

0400

0600

0800

1000

1200

Time of da

(simulated hours) Time of da

(simulated hours)

0.5

(c)

A = 0.85

A = 0.8

A = 0.75

A = 0.7

A = 0.65

A = 0.6

A = 0.55

A = 0.5

A = 0.45

A = 0.4

External signal:

sin(2St/24)/A

Increase signal

amplitude

Zero social weight (D= 0) Social information sharing (D= 0.3)

V = 0

'K = 0

V = 0

'K = 0.1

V = 0

'K = 0

V = 0

'K = 0.1

V = 0.25

'K = 0

V = 0.25

'K = 0.1

V = 0.25

'K = 0

V = 0.25

'K = 0.1

(a)

0000 0200 0400 0600 0800 1000 1200 0000 0200 0400 0600 0800 1000 1200

Time of day (simulated hours)

0.5

S(t)

External signal: S(t) = A sin(2St/24)

Signal amplitude

Nonactivated

Activated

Simulated agent

decision states:

0.5

Time of day (simulated hours)

0.5

Decision

thresholds

Figure 7. Activation model with external and social information. We modelled the activation of drones using an accumulation-to-threshold model, where evidence comes from

both an external signal as well as social information (social information is the fraction of group members already activated). (a) Model illustration. The external signal is sinusoidal to

represent circadian daily patterns, with amplitude set by the parameter A. Individual agent decision states x

follow a leaky accumulator process with noise amplitude

(equation 1). Across a simulated group of Nindividuals, each agent's decision threshold

is drawn from a normal distribution with standard deviation

. The parameter

sets the

fraction of evidence due to social information:

¼0 represents 100% use of an external signal by each agent;

¼0.3 represents 30% social evidence and 70% external signal used by

each agent. (b, c) Simulation results for zero versus nonzero values of decision noise (

) and within-group threshold differences (

), shown for (b) zero social information use and

(c) both social information and external signal used as actuation evidence. In each plot, the normalized external signal and the results for the fraction of group members activated

over time in response to different signal amplitudes Aare shown. Each 2 2 grid shows results for zero decision noise (top row), nonzero decision noise (bottom row), equal

thresholds (left column) and threshold differences among group members (right column). Simulations use a group of N¼50 agents.

L. C. Neubauer et al. / Animal Behaviour xxx (xxxx) xxx8

Please cite this article in press as: Neubauer, L. C., et al., Honey bee drones are synchronously hyperactive inside the nest, Animal Behaviour

(2023), https://doi.org/10.1016/j.anbehav.2023.05.018

threshold differences, a group with social information sharing can

synchronize the activation response (Fig. 7c). The use of social in-

formation also speeds up the onset of the activation period (how

long it takes from the ﬁrst activated individuals until approximately

all are activated), as well as shifting the time of activation onset to

relatively later (cf. Fig. 7b, c). Therefore, agents that incorporate social

information give results that more closely align with our observed

data than agents that do not incorporate social information.

DISCUSSION

Using long-term automated tracking, we followed the in-nest

behaviour of honey bee drones, examining how they move and

where they position themselves in the nest over their entire lives.

While drones spent the majority of their time immobile in the nest,

they were also synchronously hyperactive inside the nest during

afternoon periods that coincided with trips outside (Fig. 1). This

behaviour developed with age, becoming prominent from age 7

days onward, in both drone cohorts. We deﬁned the ‘shared acti-

vation period’as the time when a large fraction of drones were

active (i.e. fast-moving or taking trips outside; see Fig. 3) and found

that the onset occurred in the afternoon just after the sun was at its

peak and that the duration was longer on days with more sunshine

hours (Fig. 4,Appendix, Fig. A9). Looking at individual drones, the

onset and end of activation were synchronized: drones became

active and started/ended their outside trips at the same time as

other drones (Fig. 5). However, individual drones did not show a

strong consistency in their relative activation timing from day to

day; i.e. there were no consistent ‘early-to-activate’or ‘late-to-

depart’drones (Fig. 6a). We simulated a threshold-based model of

drone activation to demonstrate that the main observations in the

data were consistent with individual decisions using both external

and social information to decide when to activate (Fig. 7).

Daily changes in activity can be triggered by external factors

(such as weather), by internal factors (such as circadian biological

rhythms) or by social information exchange with other individuals.

It is well established that honey bee workers exhibit circadian

rhythms, which help organize roles within the colony, and can be

socially mediated (Bloch &Robinson, 2001;Eban-Rothschild &

Bloch, 2012;Medugorac &Lindauer, 1967). This likely also applies

to the timing of drone departure ﬂights. We saw that the onset of

the shared activation period occurred shortly after the time of

strongest sunlight (Appendix, Fig. A9) and that the duration of the

shared activation period was correlated with the weather condi-

tions for that day (Fig. 4). While we did observe a positive corre-

lation between the duration of the shared activation period and

daily sunshine hours (Fig. 4), is it possible that this trend could

become saturated, or even reversed, under more extreme envi-

ronmental conditions (e.g. less activation if the temperature is ‘too

hot’). Given that we did not perform manipulative experiments to

alter daily light exposure, we cannot directly distinguish between

external cues from the sun and internal triggers from biological

rhythms. However, because biological rhythms take multiple days

to establish or change (Roenneberg et al., 2003) and because

weather is highly stochastic, the absence of an increase in activity

on the poor weather day (20 June) suggests that external factors,

like weather, may play a stronger role than internal factors.

Speciﬁc experimental manipulations would be needed to test

the relative importance of internal, external and social factors for

the timing of drone hyperactivity. There is suggestive evidence that

internal factors determine ﬂight time in drones; researchers used a

ﬂight room to clock-shift drones and found that, when the drones

were moved outdoors, they retained their clock-shifted departure

time (Pfannenstiel &Koeniger, 2000). These ﬁndings, however,

were published as an abstract, and so would need to be conﬁrmed.

Ideally, such an experiment would also investigate the in-nest hy-

peractivity that we observed to potentially determine whether

clock-shifted drones could socially induce other nonshifted drones

into a hyperactive state.

The timingof individual activation did not correlate with morning

time spatial position or speed of individual drones (Fig. 6b). Con-

spicuous hyperactivity may itself be the social cue that helps to syn-

chronizeall the drones acrossthe colony, but the precisemechanisms

for information transmission and activation remain to be explicitly

tested. This could include the potential for chemical communication;

drones have been shown to exhibit age-based attraction to conspe-

ciﬁcs (Bastin et al., 2017). In our model of drone activation, we did not

consider the mechanisms of information transmission, but instead

assumed that individual drones can sense when others are activated

(equation 1). An interesting avenue for future work is to examine the

ﬁne-scale dynamics of how drones interact and respond to other

drones, including social communication mechanisms using in-

teractions and proximity among drones (Wild et al., 2021). In

conjunction with speciﬁc experimental manipulations, one could

determine how internal, external and social factors inﬂuence activa-

tion. Testing how drones coordinate their in-nest behaviour may also

provideadditional insights intothe potential for coordination outside

the nest (reviewed in Mariette et al., 2021).

Our model of drone social activation represents individuals as

leaky integrators of an external signal and social information. Using

social information can reduce uncertainty or noise in estimating

environmental states or making decisions (Ellison et al., 2016;

Srivastava &Leonard, 2014). We observed similar results in our

simulations: Fig. 7 shows that synchronized activation can occur if

independent agents have zero noise and identical activation

thresholds, or if social information is used in the presence of noise

and threshold differences. Accumulation-to-threshold models have

been applied to many different systems, with applications including

individual neuron spiking (Teeter et al., 2018), perceptual decision

making (Usher &McClelland, 2001), foraging (Bidari et al., 2022;

Davidson &El Hady, 2019) and economic or purchasing decisions

(Gluth et al., 2012;Krajbich et al., 2012). We used a leaky accumu-

lator model formulation, which previously has been applied to tasks

that require a detection of signal changes (Clifford &Ibbotson, 2002;

Glaze et al., 2015). While we represented social information sharing

with a simple fraction of activated drones, we note that future work

could incorporate additional mechanisms of information sharing

(e.g. local interactions among neighbours or nonlinear forms of

coupling; Bizyaeva et al., 2021;Zhong &Leonard, 2019).

The onset of in-nest hyperactivity aligns with important stages

in drone sexual maturation (reviewed in Koeniger et al., 2014;

Rangel &Fisher, 2019). Drone departure ﬂights ﬁrst occur at age

6e9 days of age (Reyes et al., 2019), sperm move from the testes to

the seminal vesicles at 7e8 days of age (Mackensen, 1955;

Snodgrass, 1956) and sperm viability peaks as early as 7 days of age

(Locke &Peng, 1993). While these physiological changes coincide

with the behavioural changes we observed in the nest, the extent to

which these changes are coupled remains unknown.

Drones and virgin queens have a common interest in meeting in

the most efﬁcient way possible. There are two mechanisms by

which independent parties can indirectly coordinate meetings

without direct communication: (1) restrict the spatial component

and (2) restrict the temporal component (Couzin, 2018;Dost

alkov



Spinka, 2007;Fenster et al., 1995). Drone congregation areas

were already known to restrict the spatial component, and after-

noon ﬂights restrict the temporal component. Here we show that

the restricted temporal component is also associated with a limited

period of drone hyperactivity inside the nest. Although more time

L. C. Neubauer et al. / Animal Behaviour xxx (xxxx) xxx 9

Please cite this article in press as: Neubauer, L. C., et al., Honey bee drones are synchronously hyperactive inside the nest, Animal Behaviour

(2023), https://doi.org/10.1016/j.anbehav.2023.05.018

spent outside is coupled with higher mortality (Visscher &Dukas,

1997), drones are also under selective pressure to be present in

the drone congregation area before virgin queens arrive; if they

have not arrived in time, they will miss their opportunity to mate.

These two forces contrast with one another: the need to arrive early

to overlap with virgin queens is a pressure to extend time outside,

but mortality risk limits the time spent outside. Combined, these

factors appear to be a form of stabilizing selection on the drone's

behavioural phenotype (Hansen, 1997), which results in a strong

‘on/off’activation period. In locations with high mortality outside of

the nest, such as where bee-eating birds (Meropidae) are common

(Ali &Taha, 2012;Loope, 2015), we would therefore predict a

shorter period of drone activation than in locations where mortality

is lower (Smith, 2018). In places with high outside-nest mortality,

we would still expect to see synchronized hyperactivity, but over a

shorter period. While our study did observe in-nest hyperactivity

across all tagged drones, it would also be interesting to see how

these patterns vary across colonies and environmental conditions.

A honey bee colony is an integrated superorganism, with workers

functioning as soma and dronesfunctioning as gametes (Seeley,1989;

Smith &Szathm

ary,1995). Just as the optic nerve and the testes serve

different roles within a multicellular organism, so too do the indi-

vidual bees that form a colony. It is unsurprising that drones are

referred to a s ‘lazy’(Gadagkar, 2021;von Frisch,1954); they do indeed

spend themajority of theirtime immobile at theperiphery of the nest.

However, thisbehaviour can also be viewedas adaptive: by remaining

immobile, they conserve their energy until mating time, and by

moving to the nest periphery, they prevent potential obstruction to

workers in action. Just as drones have morphological and physiolog-

ical adaptations to maximize their mating opportunities (e.g. large

eyes: Menzel et al., 1991;nohypopharyngealglands:Hrassnigg &

Crailsheim, 2005), their behaviour is also adaptive, both inside and

outside the nest. Using long-term automated tracking, we found that

drones, in contrast to their reputation, can be the most active in-

dividuals in the colony, albeit for a limited time each day. Therefore,

drones do exhibit specialized in-nest behaviour, which highlights

their role as the male gametes of the colony.

Author Contributions

Conceptualization: L.C.N., J.D.D., M.L.S. Data curation: J.D.D.

Formal analysis: L.C.N., J.D.D. Funding acquisition: T.L., I.D.C., M.L.S.

Investigation: L.C.N., M.L.S. Methodology: T.L., M.L.S. Resources: T.L.,

I.D.C. Software: B.W., D.M.D., T.L. Supervision: I.D.C., M.L.S. Visuali-

zation: L.C.N., J.D.D. Writing eoriginal draft: L.C.N., J.D.D., M.L.S.

Writing ereview &editing: L.C.N., J.D.D., B.W., D.M.D., T.L., I.D.C.,

M.L.S.

Data Availability

All data associated with this study is freely available online

through Zenodo (Smith et al., 2023). Code to reproduce the results

in the paper is available on GitHub (github.com/jacobdavidson/

bees_drones_2019data).

Declaration of Interest

None.

Acknowledgments

We thank Giovanni Galizia for providing access to his rooftop

apiary, Jan Peters for expert colony maintenance, Dagmar Olalere and

Katja Anderson for essential logistical support, Markus Miller for IT

expertise and Sinje Tigges, Christine Bauer and Jayme Weglarski for

patiently helping to tag honey bee workers and drones. This work

was supported by Heidelberger Akademie der Wissenschaften under

the WIN (Wissenschaftlichen Nachwuchs) program (M.L.S., J.D.D.),

Deutsche Forschungsgemeinschaft (German Research Foundation)

under Germany's Excellence Strategy EXC 2117-422037984 (I.D.C.),

the U.S. National Science Foundation (grant no. IOS-1355061) (J.D.D.,

I.D.C.), the Ofﬁce of Naval Research (grants no. N00014-09-1-1074

and N00014-14-1-0635) (I.D.C.), HPC-Service of ZEDAT (Freie Uni-

versit€

at Berlin) (B.W., D.M.D., T.L.), North-German Supercomputing

Alliance (B.W., D.M.D., T.L.), European Union's Horizon 2020 research

and innovation program (grant no. 824069) (D.M.D., T.L.), Klaus

Tschira Foundation (grant no. 00.300.2016) (B.W., T.L.), Andrea von

Braun Foundation and the Elsa-Neumann-Scholarship (D.M.D.),

Zukunftskolleg Mentorship Program (M.L.S., T.L.) and the Simons

Foundation Postdoctoral Fellowship of the Life Sciences Research

Foundation (M.L.S.).

Supplementary Material

Supplementary material associated with this article is available,

in the online version, at https://doi.org/10.1016/j.anbehav.2023.05.

018.

References

Ali, M. A., & Taha, E. A. (2012). Bee-eating birds (Coraciiformes: Meropidae) reduce

virgin honey bee queen survival during mating ﬂights and foraging activity of

honey bees (Apis mellifera L.). International Journal of Scientiﬁc and Engineering

Research, 3(6), 1e8.

Bastin, F., Savarit, F., Lafon, G., & Sandoz, J. C. (2017). Age-speciﬁc olfactory attraction

between western honey bee drones (Apis mellifera) and its chemical basis. PLoS

One, 12(10), 1e22. https://doi.org/10.1371/journal.pone.0185949

Beshers, S. N., & Fewell, J. H. (2001). Models of division of labor in social insects.

Annual Review of Entomology, 46(1), 413e440. https://doi.org/10.1146/

annurev.ento.46.1.413

Bidari, S., El Hady, A., Davidson, J. D., & Kilpatrick, Z. P. (2022). Stochastic dynamics

of social patch foraging decisions. Physical Review Research, 4(3). https://doi.org/

10.1103/PhysRevResearch.4.033128. Article 033128.

Bizyaeva, A., Matthews, A., Franci, A., & Leonard, N. E. (2021). Patterns of nonlinear

opinion formation on networks. arXiv:2009.13600v2 [math.OC] https://doi.org/

10.48550/arXiv.2009.13600.

Bloch, G., & Robinson, G. E. (2001). Reversal of honeybee behavioural rhythms.

Nature, 410(6832). https://doi.org/10.1038/35074183. Article 1048.

Boenisch, F., Rosemann, B., Wild, B., Dormagen, D., Wario, F., & Landgraf, T. (2018).

Tracking all members of a honey bee colony over their lifetime using learned

models of correspondence. Frontiers in Robotics and AI, 5.https://doi.org/

10.3389/frobt.2018. 00035. Article 35.

Cicciarelli, R. (2013). Drone eviction in honey bees (Apis mellifera ssp.) (M.Sc. thesis).

Cornell University https://ecommons.cornell.edu/handle/1813/34344.

Clifford, C. W. G., & Ibbotson, M. R. (2002). Fundamental mechanisms of visual

motion detection: Models, cells and functions. Progress in Neurobiology, 68(6),

409e437. https://doi.org/10.1016/S0301-0082(02)00154-5

Couzin, I. D. (2018). Synchronization: The key to effective communication in animal

collectives. Trends in Cognitive Sciences, 22(10), 844e846. https://doi.org/

10.1016/j.tics.2018.08.001

Davidson, J. D., & El Hady, A. (2019). Foraging as an evidence accumulation process.

PLoS Computational Biology, 15(7). https://doi.org/10.1371/journal.pcbi1007060.

Article e1007060.

Dost

alkov

a, I., & 

Spinka, M. (2007). Synchronization of behaviour in pairs: The role

of communication and consequences in timing. Animal Behaviour, 74(6),

173 5 e1742. https://doi.org/10.1016/j.anbehav.2007.04.014

Eban-Rothschild, A., & Bloch, G. (2012). Circadian rhythms and sleep in honey bees.

In C. G. Galizia, D. Eisenhardt, & M. Giurfa (Eds.), Honeybee neurobiology and

behavior: A tribute to Randolf Menzel (pp. 31e45). Springer. https://doi.org/

10.1007/978-94-007-2099-2.

Ellison, D., Mugler, A., Brennan, M. D., Lee, S. H., Huebner, R. J., Shamir, E. R.,

Woo, L. A., Kim, J., Amar, P., Nemenman, I., Ewald, A. J., & Levchenko, A. (2016).

Cellecell communication enhances the capacity of cell ensembles to sense

shallow gradients during morphogenesis. Proceedings of the National Academy

of Sciences of the United States of America, 113(6), E679eE688. https://doi.org/

10.1073/pnas.1516503113

Fenster, M., Kraus, S., & Rosenschein, J. S. (1995). Coordination without

communication: Experimental validation of focal point techniques. In Pro-

ceedings of the First International Confe rence on Multiagent Sy stems, ICM AS-95

(pp. 102e108).

Free, J. (1957). The food of adult drone honeybees (Apis mellifera). British Journal of

Animal Behaviour, 5,7e11.

L. C. Neubauer et al. / Animal Behaviour xxx (xxxx) xxx10

Please cite this article in press as: Neubauer, L. C., et al., Honey bee drones are synchronously hyperactive inside the nest, Animal Behaviour

(2023), https://doi.org/10.1016/j.anbehav.2023.05.018

Free, J., & Williams, I. H. (1975). Factors determining the rearing and rejection of

drones by the honeybee colony. Animal Behaviour, 23, 650e675. https://doi.org/

10.1016/0003-3472(75)90143-8

von Frisch, K. (1954). The dancing bees: An account of the life and senses of the honey

bee. Methuen.

Fukuda, H., & Ohtani, T. (1977). Survival and life span of drone honeybees. Re-

searches on Population Ecology, 19(1), 51e68. https://doi.org/10.1007/

BF02510939

Gadagkar, R. (2021). Cooperation in social insects. In T. K. Shackelford, &

V. A. Weekes-Shackelford (Eds.), Encyclopedia of evolutionary psychological sci-

ence (pp. 1462e1470). Springer International. https://doi.org/10.1007/978-3-

319-19650-3_1367.

Glaze, C. M., Kable, J. W., & Gold, J. I. (2015). Normative evidence accumulation in

unpredictable environments. eLife, 4.https://doi.org/10.7554/eLife.08825.

Article e08825.

Gluth, S., Rieskamp, J., & Büchel, C. (2012). Deciding when to decide: Time-variant

sequential sampling models explain the emergence of value-based decisions in

the human brain. Journal of Neuroscience, 32(31), 10686e10698.

Hansen, T. F. (1997). Stabilizing selection and the comparative analysis of adaptation.

Evolution, 51(5), 1341e1351. https://doi.org/10.1111/j.1558-5646.1997.tb01457.x

H€

olldobler, B., & Wilson, E. O. (2009). The superorganism: The beauty, elegance, and

strangeness of insect societies. W. W. Norton.

Hrassnigg, N., & Crailsheim, K. (2005). Differences in drone and worker physiology

in honey bees (Apis mellifera). Apidologie, 36, 255e277. https://doi.org/10.1051/

apido

Jandt, J. M., & Dornhaus, A. (2014). Bumblebee response thresholds and body size:

Does worker diversity increase colony performance? Animal Behaviour, 87,

97e106. https://doi.org/10.1016/j.anbehav.2013.10.017

Jones, J. C., Myerscough, M. R., Graham, S., & Oldroyd, B. P. (2004). Honey bee nest

thermoregulation: Diversity promotes stability. Science, 305(5682), 402e404.

https://doi.org/10.1126/science.1096340

Koeniger, G. (1986). Reproduction and mating behavior. In T. E. Rinderer (Ed.), Bee

genetics and breeding (pp. 255e280). Academic Press.

Koeniger, G., Koeniger, N., Ellis, J., & Connor, L. J. (2014). Mating biology of honey bees

(Apis mellifera). Wicwas Press.

Koeniger, N., Koeniger, G., & Pechhacker, H. (2005). The nearer the better? Drones

(Apis mellifera) prefer nearer drone congregation areas. Insectes Sociaux, 52(1),

31e35. https://doi.org/10.1007/s00040-004-0763-z

Koeniger, N., Koeniger, G., Tingek, S., Mardan, M., & Rinderer, T. E. (1988). Repro-

ductive isolation by different time of drone ﬂight between Apis cerana (Fabricus,

1793) and Apis vechti (Maa, 1953). Apidologie, 19(1), 103e106.

Kovac, H., Stabentheiner, A., & Brodschneider, R. (2009). Contribution of honeybee

drones of different age to colonial thermoregulation. Apidologie, 40(1), 82e95.

https://doi.org/10.1051/apido/2008069

Krajbich, I., Lu, D., Camerer, C., & Rangel, A. (2012). The attentional drift-diffusion

model extends to simple purchasing decisions. Frontiers in Psychology, 3.

Article 193.

Locke, S. J., & Peng, Y.-S. (1993). The effects of drone age, semen storage and

contamination on semen quality in the honey bee (Apis mellifera). Physio-

logical Entomology, 18(2), 144e148. h t t p s : / / doi . o r g / 10 .1111 / j .1 3 65 -

3032.1993.tb00461.x

Loope, K. J. (2015). Reproductiveconﬂicts and signal evolution in social wasps and bees

(Ph.D. thesis). Cornell University https://ecommons.cornell.edu/handle/1813/

41121.

Loper, G. M., Wolf, W. W., & Taylor, O. R., Jr. (1992). Honey bee drone ﬂyways and

congregation areas: Radar observations. Journal of the Kansas Entomological

Society, 65(3), 223e230. https://www.jstor.org/stable/25085360.

Mackensen, O. (1955). Experiments in the technique of artiﬁcial insemination of

queen bees. Journal of Economic Entomology, 48(4), 418e421. https://doi.org/

10.1093/jee/48.4.418

Mariette, J., Carcaud, J., & Sandoz, J. C. (2021). The neuroethology of olfactory sex

communication in the honeybee Apis mellifera L. Cell and Tissue Research,

383(1), 177e194. https://doi.org/10.1007/s00441-020-03401-8

Mattila, H. R., Rios, D., Walker-Sperling, V. E., Roeselers, G., & Newton, I. L. (2012).

Characterization of the active microbiotas associated with honey bees reveals

healthier and broader communities when colonies are genetically diverse. PLoS

One, 7(3). https://doi.org/10.1371/journal.pone.0032962. Article e32962.

Mattila, H. R., & Seeley, T. D. (2007). Genetic diversity in honey bee colonies en-

hances productivity and ﬁtness. Science, 317(5836), 362e364. https://doi.org/

10.1126/science.143046

Medugorac, I., & Lindauer, M. (1967). Das Zeitged€

achtnis der Bienen unter dem

Einﬂuß von Narkose und von sozialen Zeitgebern. Zeitschrift für Vergleichende

Physiologie, 55(4), 450e474. https://doi.org/10.1007/BF00302625

Menzel, J. G., Wunderer, H., & Stavenga, D. G. (1991). Functional morphology of the

divided compound eye of the honeybee drone (Apis mellifera). Tissue and Cell,

23(4), 525e535. https://doi.org/10.1016/0040-8166(91)90010-Q

Mizumoto, N., & Dobata, S. (2019). Adaptive switch to sexually dimorphic move-

ments by partner-seeking termites. Science Advances, 5(6). https://doi.org/

10.1126/sciadv.aau6108. Article eaau6108.

Morse, R. A., Strang, G. E., & Nowakowski, J. (1967). Fall death rates of drone honey

bees. Journal of Economic Entomology, 60(5), 1198e1202. https://doi.org/

10.1093/jee/605.1198

Neves, E. F., Faita, M. R., de Oliveira Gaia, L., Vieira Alves Júnior, V. V., & Antonialli-

Junior, W. F. (2011). Inﬂuence of climate factors on ﬂight activity of drones of

Apis mellifera (Hymenoptera: Apidae). Sociobiology, 57(1), 107e114113 .

Oertel, E. (1956). Observations on the ﬂight of drone honey bees. Annals of the

Entomological Society of America, 49(5), 497e500. https://doi.org/10.1093/aesa/

49.5.497

Oldroyd, B. P., & Wongsiri, S. (2006). Asian honey bees: Biology, conservation, and

human interactions. Harvard University Press.

O'Shea-Wheller, T. A., Hunt, E. R., & Sasaki, T. (2021). Functional heterogeneity in

superorganisms: Emerging trends and concepts. Annals of the Entomological

Society of America, 114(5), 562e574. https://doi.org/10.1093/aesa/saaa039

Paxton, R. J. (2005). Male mating behavior and mating systems of bees: An over-

view. Apidologie, 36,145e156. https://doi.org/10.1051/apido

Peitsch, D., Fietz, A., Hertel, H., de Souza, J., Ventura, D. F., & Menzel, R. (1992). The

spectral input systems of hymenopteran insects and their receptor-based colour

vision. Journal of Comparative Physiology A, 170(1), 23e40. https://doi.org/

10.1007/BF00190398

Pfannenstiel, E., & Koeniger, N. (2000). Endogenous control of drone ﬂight times

(Apis mellifera L.). Apidologie, 31(5), 641e642.

Rangel, J., & Fisher, A. (2019). Factors affecting the reproductive health of honey bee

(Apis mellifera) drones: A review. Apidologie, 50(6), 759e778. https://doi.org/

10.1007/s13592-019-00684-x

Reyes, M., Crauser, D., Prado, A., & Le Conte, Y. (2019). Flight activity of honey bee

(Apis mellifera) drones. Apidologie, 50(5), 669e680. https://doi.org/10.1007/

s13592-019-00677-w

Roenneberg, T., Daan, S., & Merrow, M. (2003). The art of entrainment. Journal of

Biological Rhythms, 18(3), 183e194. https://doi.org/10.1177/

0748730403018003001

Ruttner, F. (1966). The life and ﬂight activity of drones. Bee World, 47(3), 93e100.

https://doi.org/10.1080/0005772X.1966.11097111

Ruttner, F., & Ruttner, H. (1966). Untersuchungen über die Flugaktivit€

at und das

Paarungsverhalten der Drohnen. iii. Flugweite und Flugrichtung der Drohnen.

Zeitschrift für Bienenforschung, 8, 332e354.

Ruttner, H., & Ruttner, F. (1972). Untersuchungen über die Flugaktivit€

at und das

Paarungsverhalten der Drohnen. v. eDrohnensammelpl€

atze und Paar-

ungsdistanz. Apidologie, 3(3), 203e232.

Seeley, T. D. (1989). The honey bee colony as a superorganism. American Scientist,

77(6), 546e553. http://www.jstor.org/stable/27856005.

Seeley, T. D. (1995). The wisdom of the hive: The social physiology of honey bee col-

onies. Harvard University Press. https://doi.org/10.1016/j.desal.2010.03.003

Seeley, T. D. (2010). Honeybee democracy. Princeton University Press. https://doi.org/

10.1515/9781400835959

Seeley, T. D. (2019). The lives of bees: The untold story of the honey bee in the wild.

Princeton University Press.

Smith, P. L. (2000). Stochastic dynamic models of response time and accuracy: A

foundational primer. Journal of Mathematical Psychology, 44(3), 408e463.

https://doi.org/10.1006/jmps.1999.1260

Smith, M. L. (2018). Queenless honey bees build infrastructure for direct repro-

duction until their new queen proves her worth. Evolution, 72(12), 2810e2817.

https://doi.org/10.1111/evo.13628

Smith, M. L., Davidson, J. D., Wild, B., Dormagen, D. M., Landgraf, T., & Couzin, I. D.

(2022). Behavioral variation across the days and lives of honey bees. iScience,

25(9). https://doi.org/10.1016/j.isci.2022.104842. Article 104842.

Smith, M. L., Davidson, J. D., Wild, B., Dormagen, D. M., Landgraf, T., & Couzin, I. D.

(2023). Honeybee lifetime tracking data 2019. Zenodo.https://doi.org/10.5281/

zenodo.7298798.

Smith, M. L., Ostwald, M. M., & Seeley, T. D. (2016). Honey bee sociometry: Tracking

honey bee colonies and their nest contents from colony founding until death.

Insectes Sociaux, 63, 553e563. https://doi.org/10.1007/s00040-016-0499-6

Smith, J. M., & Szathm

ary, E. (1995). The major transitions in evolution. Oxford

University Press.

Snodgrass, R. E. (1956). Anatomy of the honey bee. Cornell University Press. https://

doi.org/10.5962/bhl.title.87234

Srivastava, V., & Leonard, N. E. (2014). Collective decision-making in ideal networks:

The speedeaccuracy tradeoff. IEEE Transactions on Control of Network Systems,

1(1), 121e132. https://doi.org/10.1109/TCNS.2014.2310271

Taber, S., 3rd (1964). Factors inﬂuencing the circadian ﬂight rhythm of drone honey

bees. Annals of the Entomological Society of America, 57(6), 769e775. https://

doi.org/10.1093/aesa/57.6.769

Tarpy, D. R. (2003). Genetic diversity within honeybee colonies prevents severe

infections and promotes colony growth. Proceedings of the Royal Society B:

Biological Sciences, 270(1510), 99e103. https://doi.org/10.1098/rspb.2002.2199

Teeter, C., Iyer, R., Menon, V., Gouwens, N., Feng, D., Berg, J., Szafer, A., Cain, N.,

Zeng, H., Hawrylycz, M., Koch, C., & Mihalas, S. (2018). Generalized leaky

integrate-and-ﬁre models classify multiple neuron types. Nature Communica-

tions, 9(1). https://doi.org/10.1038/s41467-017-02717-4. Article 709.

Tschinkel, W. R. (2006). The ﬁre ants. Belknap Press of Harvard University Press.

Ulrich, Y., Kawakatsu, M., Tokita, C. K., Saragosti, J., Chandra, V., Tarnita, C. E., &

Kronauer, D. J. C. (2021). Response thresholds alone cannot explain empirical

patterns of division of labor in social insects. PLoS Biology, 19(6). https://doi.org/

10.1371/journal.pbio.3001269. Article e3001269.

Usher, M., & McClelland, J. L. (2001). The time course of perceptual choice: The

leaky, competing accumulator model. Psychological Review, 108(3), 550e592.

https://doi.org/10.1037/0033-295X.108.3.550

Visscher, P., & Dukas, R. (1997). Survivorship of foraging honey bees. Insectes

Sociaux, 44(1), 1e5. https://doi.org/10.1007/s000400050017

Wario, F., Wild, B., Couvillon, M. J., Rojas, R., & Landgraf, T. (2015). Automatic

methods for long-term tracking and the detection and decoding of

L. C. Neubauer et al. / Animal Behaviour xxx (xxxx) xxx 11

Please cite this article in press as: Neubauer, L. C., et al., Honey bee drones are synchronously hyperactive inside the nest, Animal Behaviour

(2023), https://doi.org/10.1016/j.anbehav.2023.05.018

communication dances in honeybees. Frontiers in Ecology and Evolution, 3.

https://doi.org/10.3389/fevo. 2015.00103. Article 103.

Weismann, A. (1893). The germ-plasm: A theory of heredity. Scribner.

Wharton, K. E., Dyer, F. C., & Getty, T. (2008). Male elimination in the honeybee.

Behavioral Ecology, 19(6), 1075e1079. https://doi.org/10.1093/beheco/arn108

Wild, B., Dormagen, D. M., Zachariae, A., Smith, M. L., Traynor, K. S., Brockmann, D.,

Couzin, I. D., & Landgraf, T. (2021). Social networks predict the life and death of

honey bees. Nature Communications, 12(1), 1e12. https://doi.org/10.1038/

s41467-021-21212-5

Witherell, P. C. (1971). Duration of ﬂight and of interﬂight time of drone honey bees,

Apis mellifera.Annals of the Entomological Society of America, 64(3), 609e612.

https://doi.org/10.1093/aesa/64.3.609

Wongsiri, S., Lekprayoon, C., Thapa, R., Thirakupt, K., Rinderer, T. E., Sylvester, H. A.,

Oldroyd, B. P., & Booncham, U. (1997). Comparative biology of Apis andreniformis

and Apis ﬂorea in Thailand. Bee World, 78(1), 23e35. https://doi.org/10.1080/

0005772X.1997.11099328

Woodgate, J. L., Makinson, J. C., Rossi, N., Lim, K. S., Reynolds, A. M., Rawlings, C. J., &

Chittka, L. (2021). Harmonic radar tracking reveals that honeybee drones

navigate between multiple aerial leks. iScience, 24(6). https://doi.org/10.1016/

j.isci.2021.102499. Article 102499.

Zhong, Y. D., & Leonard, N. E. (2019). A continuous threshold model of cascade

dynamics. In Proceedings of the 2019 IEEE 58th Conference on Decision and

Control (CDC), Nice, France, 11e13 December 2019 (pp. 1704e1709). IEEE. https://

doi.org/10.1109/CDC40024.2019.9029844.

Appendix

8 June 9 June 10 June 11 June 12 June 13 June 14 June 15 June 16 June 17 June 18 June

Age 1 Age 2 Age 3 Age 4 Age 5 Age 6 Age 7 Age 8 Age 9 Age 10 Age 11

Age 1 Age 2 Age 3 Age 4 Age 5

Age 2 Age 3 Age 4 Age 5 Age 6

Age 2

Age 1

Age 0

Age 0 Age 3 Age 4 Age 5 Age 6

Age 6 Age 7 Age 8 Age 9 Age 10 Age 11

Drones, cohort 1

19 June 20 June 21 June 22 June 23 June 24 June 25 June 26 June 27 June 28 June 29 June

Age 12 Age 13 Age 14 Age 15 Age 16 Age 17 Age 18 Age 19 Age 20 Age 21 Age 22

Drones, cohort 1

Age 12 Age 13 Age 14 Age 15 Age 16 Age 17 Age 18 Age 19 Age 20 Age 21 Age 22

Workers, cohort 1

Age 7 Age 8

Age 9

Age 10 Age 11 Age 12 Age 13 Age 14 Age 15 Age 16 Age 17

Drones, cohort 2

Age 7 Age 8

Age 9

Age 10 Age 11 Age 12 Age 13 Age 14 Age 15 Age 16 Age 17

Workers, cohort 2

Drones, cohort 2

Workers, cohort 2

Workers, cohort 1

Honey

Capped brood

Young brood

Empty comb

Pollen stores

Wooden frames

Festoon

Blank space

20 cm

Figure A1. Two-dimensional spatial histograms showing the locations and ages of workers and drones on each day. Analogous results are shown in Fig. 2 using a subset of the

observation days shown here. The top row shows nest contents over the observation period. For days where comb measurements were not taken (14, 15, 16 and 29 June), the comb

was not known exactly, so both of the nearest measurement days are shown overlayed, with the transparency weighted proportionally to the time from the measurement day (i.e.

less transparency for the closer measurement day).

L. C. Neubauer et al. / Animal Behaviour xxx (xxxx) xxx12

Please cite this article in press as: Neubauer, L. C., et al., Honey bee drones are synchronously hyperactive inside the nest, Animal Behaviour

(2023), https://doi.org/10.1016/j.anbehav.2023.05.018

6.25

12.5

18.75

31.25

37.5

43.75

56.25

62.5

68.75

31.25

37.5

43.75

56.25

62.5

68.75

Figure A2. Exit distance (cm) for different locations in the observation hive. Note that the exit is located at the lower right of the front side of the observation hive (lower right of

images) and that crossings from one side to the other are possible on the frame borders. To estimate when bees exited the nest, we used a threshold of 31.25 cm on median exit

distance in the 5 min bins (see Methods for details on this algorithm); this value is made bold in the ﬁgure.

10 15 20 25 30

of June 2019

Avg. dist. from frame

centre in observation

hive (cm)

Drones, cohort 1

Drones, cohort 2

Workers, cohort 1

Workers, cohort 2

Figure A3. Average distance from frame centre. The observation hive consisted of three frames, and each frame had two sides (see Fig. 2,Fig. A1). Shown is the average distance

from centre of the current frame, comparing drone versus worker over time.

L. C. Neubauer et al. / Animal Behaviour xxx (xxxx) xxx 13

Please cite this article in press as: Neubauer, L. C., et al., Honey bee drones are synchronously hyperactive inside the nest, Animal Behaviour

(2023), https://doi.org/10.1016/j.anbehav.2023.05.018

of June 2019

14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29

1400

1500

1600

1700

1800 (a)

Shared activation

period (time of day, hours)

sact = 0.5, f = 0.25

sact = 0.4

sact = 0.6

1400

1500

1600

1700

1800 (b)

sact = 0.5, f = 0.25

f = 0.2

f = 0.3

14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29

Figure A4. Sensitivity of shared activation period deﬁnition to parameter values. In the main analysis, the value of speed threshold used was s

act

¼0.5 and the fraction of drones

f¼0.25. Shown is how the duration of the shared activation period s

act

depends on (a) changing the speed threshold and (b) changing the fraction of drones threshold f. While the

deﬁnition is not sensitive to the speed threshold, individual days have patterns that are sensitive to the value of f, in particular, using a lower drone fraction threshold deﬁnes a

longer shared activation period for 21 June, because it then would extend to include the whole time until the small ‘spike’in activity near 1700 hours, which followed a period of low

speed/inside. The parameters of s

act

¼0.5 and f¼0.25 used in the main analysis were chosen, by visual inspection, to be representative of capturing the observed quick start and

end of the shared activation period (see Fig. 3a).

–60 –50 –40 –30 –20 –10

Time before first exit associated

with shared activation

eriod (min)

0.25

0.5

0.75

Median speed (cm/s)

All

Day of June 2019

0 1 2 3 45

Number of trips

0.1

0.2

0.3

Probability density

010 20 30 40 50 60 70 80

Time per trip (min)

0.01

0.02

0.03

(a)

(b)

Figure A5. Individual drone outside trips and speed before exiting. (a) The distributions of the number of trips per drone (left) and the time per trip (right) for drones during the

shared activation periods. Distributions include data across all days with a shared activation period. (b) Median speed of drones, in the time period preceding their ﬁrst exit

associated with the shared activation period (ﬁrst exit time was deﬁned as the ﬁrst time bin tstarting from 1 h preceding the shared activation period where the drone was outside

the hive).

L. C. Neubauer et al. / Animal Behaviour xxx (xxxx) xxx14

Please cite this article in press as: Neubauer, L. C., et al., Honey bee drones are synchronously hyperactive inside the nest, Animal Behaviour

(2023), https://doi.org/10.1016/j.anbehav.2023.05.018

Speed of all workers

Day of June 2019

(a)

0000

0100

0200

0300

0400

0500

0600

0700

0800

0900

1000

1100

1200

1300

1400

1500

1600

1700

1800

1900

2000

2100

2200

2300

Speed of workers on exit frame

(b)

0000

0100

0200

0300

0400

0500

0600

0700

0800

0900

1000

1100

1200

1300

1400

1500

1600

1700

1800

1900

2000

2100

2200

2300

Time of da

(hours)

Figure A6. Average movement of workers during each day. The average speed of workers over time each day, calculated as the mean of the median 5 min bin speed of workers. Data

are plotted analogous to Fig. 3b, with the shared activation period of the drones highlighted in red and the range for each day at 0e2.5 cm/s (this is the same scale as Fig. 3b). Shown

are averages for (a) all workers and (b) workers whose fraction of time observed on the exit frame in a given 5 min time bin was at least 50%.

Time of da

(hours)

Worker cohorts 1 and 2:

Median speed

Day of June 2019

1200 1300 1400 1500 1600 1700 1800

Speed

(cm/s)

1.75

1.5

1.25

0.75

0.5

0.25

(a) 13

Worker cohorts 1 and 2:

Out of hive

1200 1300 1400 1500 1600 1700 1800

(b) Comb

measurement

period

Shared

activation

start/end

In hive (grey)

Out of hive

Figure A7. Worker bee cohorts: individual activity patterns. Analogousplots in Fig. 5 but showing worker bee cohorts 1 and 2 (who had the same birth dates, respectively,as drone cohorts 1 and

2). In the raster plots, each row represents an individual bee on a particular day,showing (a) speed and ( b) trips outside the hive, in 5 min bins (see Methods for data processing details). As in Fig. 5,

white areas are when cameras were temporarily blocked to transcribe nest contents, and red lines denote the start/end of the shared activation period ofthedrones(seealsoFig. 3b, c).

L. C. Neubauer et al. / Animal Behaviour xxx (xxxx) xxx 15

Please cite this article in press as: Neubauer, L. C., et al., Honey bee drones are synchronously hyperactive inside the nest, Animal Behaviour

(2023), https://doi.org/10.1016/j.anbehav.2023.05.018

MeteoBlue-Litzelstetten

DWD-Konstanz

DWD-Singen

DWD-Stochach

DWD-Guttingen

DWD-Friedrichshafen-

Unterraderach

Nearby weather stations

Other regional stations

Average

temperature

(ºC)

Sunshine

duration (h)

Precipitation

(mm)

1010

1015

1020

Pressure

(hPA)

Average wind

speed (km/h)

6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

of June 2019

Vapor pressure

deficit (hPa)

Figure A8. Daily weather for observation days. Nearby and regional weather data for observation days, showing average temperature, sunshine duration, total precipitation, average

atmospheric pressure, average wind speed and average vapor pressure deﬁcit for each day. Not all measurements were available for all weather stations. Recorded values of

precipitation and wind speed sometimes differed for the different weather stations, but other measurements were similar. Nearby station Konstanz-Litzelstetten data are from

MeteoBlue (2.2 km from hive), and other data are open data from the Deutsche Wetterdienst (DWD). Station Konstanz was the closest at 1.6 km from the hive.

L. C. Neubauer et al. / Animal Behaviour xxx (xxxx) xxx16

Please cite this article in press as: Neubauer, L. C., et al., Honey bee drones are synchronously hyperactive inside the nest, Animal Behaviour

(2023), https://doi.org/10.1016/j.anbehav.2023.05.018

(b) Start of shared activation period compared to weather

1200

1300

1400

1500

1600

1700

1800

Max. hour of

normalized tem

erature

1200

1300

1400

1500

1600

1700

1800

Start of shared activation

period (time of day, hours)

r=–0.44 (P=0.101)

1200

1300

1400

1500

1600

1700

1800

Max. hour of

normalized shortwave radiation

r=–0.033 (P=0.907)

1200

1300

1400

1500

1600

1700

1800

Max. hour of

normalized va

ressure deficit

r=–0.282 (P=0.308)

Shared activation

after max. hour

Shared activation

before max. hour

Shared activation

at max. hour

(a)

0000

0100

0200

0300

0400

0500

0600

0700

0800

0900

1000

1100

1200

1300

1400

1500

1600

1700

1800

1900

2000

2100

2200

2300

Time of day (hours)

0000

0100

0200

0300

0400

0500

0600

0700

0800

0900

1000

1100

1200

1300

1400

1500

1600

1700

1800

1900

2000

2100

2200

2300

0000

0100

0200

0300

0400

0500

0600

0700

0800

0900

1000

1100

1200

1300

1400

1500

1600

1700

1800

1900

2000

2100

2200

2300

Day of June 2019

Normalized temperature (7.3–32.9 °C) Normalized shortwave radiation (0–881.1 W/m2) Normalized vapor pressure deficit (0–27.4 hPa)

Figure A9. Hourly weather and start of activation period. (a) Detailed hourly temperature, sunshine intensity and vapor pressure deﬁcit recorded at Konstanz-Litzelstetten

(MeteoBlue, 2.2 km from the hive) for selected observation days. These data are plotted analogous to speed and number of drones outside the nest in Fig. 3b, c, with the

shared activation period highlighted in red. Points show the time of day for the maximum value of each quantity. (b) Start of the shared activation period (Yaxis) plotted as a

function of time of peak weather quantities (Xaxis), using the peak times shown in (a). Pearson correlations (r) and Pvalues are shown on each plot. The dashed lines show the

expected trend if the start of the shared activation period coincided with the maximum hour of each weather quantity. See also Fig. 4 for a comparison of daily weather averages for

nearby stations with the duration of the shared activation period.

L. C. Neubauer et al. / Animal Behaviour xxx (xxxx) xxx 17

Please cite this article in press as: Neubauer, L. C., et al., Honey bee drones are synchronously hyperactive inside the nest, Animal Behaviour

(2023), https://doi.org/10.1016/j.anbehav.2023.05.018

Variability of wing morphology of Apis mellifera in Ukraine

Article

Full-text available

Jan 2025

The morphological characteristics of drone wings were studied at apiaries in Khmelnytskyi, Lviv, Zhy-tomyr, Kyiv, Sumy, and Kharkiv oblasts of Ukraine. Using the classical morphometry, the mean values of 8 traits (indices: Ci, Dbi, Disc.sh, Pci, Ri (traditional), Ci.3, Ci.2.1, Ci.2.2 (proposed by the authors)) were obtained for individual clusters of drone wings from the studied colonies. The phenotypic identity of 299 drone wing clusters was determined, 114 of which were classified as belonging to the subspecies A.m.carnica, 107 to the population of Ukrainian honeybees (proposed indice: "UkrBee"), and 78 to the population of Ukrainian steppe honeybees (proposed indice: "UkrStep"). An additional allocation of individual wings for each cluster was carried out among three reference groups (templates), one of which belonged to the subspecies Apis mellifera carnica, while the other two represented the local populations with subspecies affiliation not reliably established. Based on this classification, the degrees of similarity (DS) of the larger and smaller parts of the wings in each studied cluster to the main and "admixture" phenotypes were determined. The authors propose using the degree of similarity (DS) for the smaller portion of the colony cluster sample as an indicator of hybridization and designate it as the degree of hybridization (DH) of queen genomes in the region responsible for the wing phenotype. To determine the relationships between Mahalanobis distances (MD), classification traits: Ci, Disc.sh on the one hand, and hybridization type and DS on the other, a variance analysis was performed. The obtained statistically significant relationships allowed analyzing the morphometric data in order to assess the type and degree of similarity among drone wing phenotypes. A method of determining the total genome hybridization of drone genomes in the part responsible for wing morphology, for individual colonies of honeybee populations in Ukraine has been proposed.

Factors affecting heat resilience of drone honey bees (Apis mellifera) and their sperm

Article

Full-text available

Feb 2025
PLOS ONE

Extreme temperatures associated with climate change are expected to impact the physiology and fertility of a variety of insects, including honey bees. Most previous work on this topic has focused on female honey bees (workers and queens), and comparatively little research has investigated how heat exposure affects males (drones). To address this gap, we tested body mass, viral infections, and population origin as predictors of drone survival and sperm viability in a series of heat challenge assays. We found that individual body mass was highly influential, with heavier drones being more likely to survive a heat challenge (4 h at 42°C) than smaller drones. In a separate experiment, we compared the survival of Northern California and Southern California drones in response to the same heat challenge (4 h at 42°C), and found that Southern Californian drones ― which are enriched for African ancestry ― were more likely to survive a heat challenge than drones originating from Northern California. To avoid survivor bias, we conducted sperm heat challenges using in vitro assays and found remarkable variation in sperm heat resilience among drones sourced from different commercial beekeeping operations, with some exhibiting no reduction in sperm viability after heat challenge and others exhibiting a 75% reduction in sperm viability. Further investigating potential causal factors for such variation, we found no association between drone mass and viability of sperm in in vitro sperm heat challenge assays, but virus inoculation (with Israeli acute paralysis virus) exacerbated the negative effect of heat on sperm viability. These experiments establish a vital framework for understanding the importance of population origin and comorbidities for drone heat sensitivity.

Non-Intrusive System for Honeybee Recognition Based on Audio Signals and Maximum Likelihood Classification by Autoencoder

Article

Full-text available

Aug 2024
SENSORS-BASEL

Artificial intelligence and Internet of Things are playing an increasingly important role in monitoring beehives. In this paper, we propose a method for automatic recognition of honeybee type by analyzing the sound generated by worker bees and drone bees during their flight close to an entrance to a beehive. We conducted a wide comparative study to determine the most effective preprocessing of audio signals for the detection problem. We compared the results for several different methods for signal representation in the frequency domain, including mel-frequency cepstral coefficients (MFCCs), gammatone cepstral coefficients (GTCCs), the multiple signal classification method (MUSIC) and parametric estimation of power spectral density (PSD) by the Burg algorithm. The coefficients serve as inputs for an autoencoder neural network to discriminate drone bees from worker bees. The classification is based on the reconstruction error of the signal representations produced by the autoencoder. We propose a novel approach to class separation by the autoencoder neural network with various thresholds between decision areas, including the maximum likelihood threshold for the reconstruction error. By classifying real-life signals, we demonstrated that it is possible to differentiate drone bees and worker bees based solely on audio signals. The attained level of detection accuracy enables the creation of an efficient automatic system for beekeepers.

Sex and caste effects on the vibrational sensitivity in honey bees (Apis mellifera)

Article

Full-text available

Jun 2024

In the darkness of their nests, most social insect species communicate relying on chemical, tactile, electrical and mechanical signals. In honeybees, vibrational signals play a role not only in worker communication but also in communication among virgin queens in the process of swarming. Whereas the sensitivity to vibrations has been well studied in worker bees, vibrational sensitivity of queens and drones has never been investigated. We therefore studied the sensitivity to vibrations comparatively in workers, virgin and laying queens and in drones, focusing on the frequency range mainly used by virgin queens (350–500 Hz). Bees were tested behaviorally for responses to pulses of substrate borne vibrations in arenas placed on vibrational exciters. Vibrational amplitudes were measured using (LDV-calibrated) accelerometers. Real stimuli and sham stimuli were presented in a pseudorandom order. The threshold was defined as the lowest tested amplitude at which significantly more behavioral reactions were shown to real stimuli than to sham stimuli. Workers and virgin queens respond to amplitudes down to 0.25 m/s ² . The thresholds of laying queens are substantially higher and they respond to minimum amplitudes of 0.55 m/s ² . Drones show responses to amplitudes down to 0.6 m/s ² . We conclude that sex and caste have effects on vibrational sensitivity in honeybees: virgin queens are significantly more sensitive than laying queens, indicating that a high sensitivity is crucial for survival during the process of swarming; workers are likewise highly sensitive to execute efficient everyday work; drones are significantly the least sensitive.

To house or oust: Honey bee (Apis mellifera) colonies can evaluate and evict drones of low quality

Article

Full-text available

Apr 2024
BEHAV ECOL SOCIOBIOL

Across the animal kingdom, males advertise their quality to potential mates. Males of low reproductive quality, such as those that are sick, may be excluded from mating. In eusocial species, there is some evidence that reproductive females gauge the quality of their mates. However, males often spend much more time with non-reproductive females when being raised or when returning from unsuccessful mating flights. Do non-reproductive workers evaluate the quality of male reproductives? Here we address this question using male honey bees (Apis mellifera), called drones, as a model. We generated immune-challenged drones by injecting them with lipopolysaccharide and tested: 1) do workers evict immune-challenged drones from their colony, 2) do cuticular hydrocarbon (CHC) profiles, body size, or mass change when drones are immune-challenged, and 3) are these changes used by workers to exclude low quality males from the colony? We found that an immune challenge causes changes in CHC profiles of drones and reduces their body mass. Workers selectively evict small and immune-challenged drones who, themselves, do not self-evict. This work demonstrates that some eusocial males undergo an additional layer of scrutiny prior to mating mediated by the non-reproductive worker caste. Significance statement Males of some species must advertise their quality to mates but, in the case of eusocial species, must they also advertise their quality to nestmates? By manipulating honey bee male quality, we found that small and immune-challenged drones are evicted from colonies overnight. Workers may not use a drone’s cuticular hydrocarbon profile to make this assessment. This is a new example of social immunity expressed against adult males and an example of worker involvement in reproductive decisions.

A quantitative Apis mellifera hazard and risk assessment model (AMHRA) illustrated with the insecticide sulfoxaflor: sulfoxaflor environmental science review part VI

Article

Mar 2025
J TOXICOL ENV HEAL B

Ecological comparison of native (Apis mellifera mellifera) and hybrid (Buckfast) honeybee drones in southwestern Sweden indicates local adaptation

Article

Full-text available

Aug 2024
PLOS ONE

Honeybee drones’ only known task is to mate with a virgin queen. Apart from their mating behaviour, their ecology has been little studied, especially in comparison to honeybee females. Previous knowledge is primarily based on short-term direct observations at single experimental hives, rarely, if ever, addressing the effect of drones’ genetic origin. Here, Radio Frequency Identification Technology was utilised to gather drone and worker bee lifetime data of Apis mellifera mellifera and Apis mellifera x (hybrid Buckfast) colonies over one mating season (spring and summer) with the ultimate goal to investigate differences at subspecies level. This technique enabled continuous monitoring of tagged bees at the hive entrance and recording of individuals’ movement directions. The results confirmed that spring-born drones survive longer than summer-born drones and that they generally live longer than worker bees. Drones’ peak activity occurred in the afternoon while worker bees showed more even activity levels throughout the day. Earlier orientation flights than usually reported for drones were observed. In summer, mating flights were practiced before reaching sexual maturity (at 12 days of age). Differences were found between Apis m. mellifera and Buckfast drones, where Apis m. mellifera showed later drone production in spring, but significantly earlier first activities outside the hive in summer and a later peak in diurnal activity. Additionally, Apis m. mellifera flew more in higher light intensities and windy conditions and performed significantly longer flights than Buckfast drones. The observed differences in drone ecology indicate the existence of a local adaptation of the native subspecies Apis m. mellifera to environmental conditions in southwestern Sweden.

How honeybees respond to heat stress from the individual to colony level

Article

Full-text available

Oct 2023
J R SOC INTERFACE

A honey bee colony functions as an integrated collective, with individuals coordinating their behaviour to adapt and respond to unexpected disturbances. Nest homeostasis is critical for colony function; when ambient temperatures increase, individuals switch to thermoregulatory roles to cool the nest, such as fanning and water collection. While prior work has focused on bees engaged in specific behaviours, less is known about how responses are coordinated at the colony level, and how previous tasks predict behavioural changes during a heat stress. Using BeesBook automated tracking, we follow thousands of individuals during an experimentally induced heat stress, and analyse their behavioural changes from the individual to colony level. We show that heat stress causes an overall increase in activity levels and a spatial reorganization of bees away from the brood area. Using a generalized framework to analyse individual behaviour, we find that individuals differ in their response to heat stress, which depends on their prior behaviour and correlates with age. Examining the correlation of behavioural metrics over time suggests that heat stress perturbation does not have a long-lasting effect on an individual’s future behaviour. These results demonstrate how thousands of individuals within a colony change their behaviour to achieve a coordinated response to an environmental disturbance.

Stochastic dynamics of social patch foraging decisions

Article

Full-text available

Aug 2022

Animals typically forage in groups. Social foraging can help animals avoid predation and decrease their uncertainty about the richness of food resources. Despite this, theoretical mechanistic models of patch foraging have overwhelmingly focused on the behavior of single foragers. In this study, we develop a mechanistic model that accounts for the behavior of individuals foraging together and departing food patches following an evidence accumulation process. Each individual's belief about patch quality is represented by a stochastically accumulating variable, which is coupled to another's belief to represent the transfer of information. We consider a cohesive group, and model information sharing by considering both intermittent pulsatile coupling (only communicate decision to leave) and continuous diffusive coupling (communicate throughout the deliberation process). Groups employing pulsatile coupling can obtain higher foraging efficiency, which depends more strongly on the coupling parameter compared to those using diffusive coupling. Conversely, groups using diffusive coupling are more robust to changes and heterogeneities in belief weighting and departure criteria. Efficiency is measured by a reward rate function that balances the amount of energy accumulated against the time spent in a patch, computed by solving an ordered first passage time problem for the patch departures of each individual. Using synthetic departure time data, we can distinguish between the two modes of communication and identify the model parameters. Our model establishes a social patch foraging framework to identify deliberative decision strategies and forms of social communication, and to allow model fitting to field data from foraging animal groups.

Behavioral variation across the days and lives of honey bees

Article

Full-text available

Aug 2022

In honey bee colonies, workers generally change tasks with age (from brood care, to nest work, to foraging). While these trends are well-established, our understanding of how individuals distribute tasks during a day, and how individuals differ in their lifetime behavioral trajectories, is limited. Here, we use automated tracking to obtain long-term data on 4,100+ bees tracked continuously at 3Hz, across an entire summer, and use behavioral metrics to compare behavior at different timescales. Considering single days, we describe how bees differ in space use, detection, and movement. Analyzing the behavior exhibited across their entire lives, we find consistent inter-individual differences in the movement characteristics of individuals. Bees also differ in how quickly they transition through behavioral space to ultimately become foragers, with fast-transitioning bees living the shortest lives. Our analysis framework provides a quantitative approach to describe individual behavioral variation within a colony from single days to entire lifetimes.

Response thresholds alone cannot explain empirical patterns of division of labor in social insects

Article

Full-text available

Jun 2021
PLOS BIOL

The effects of heterogeneity in group composition remain a major hurdle to our understanding of collective behavior across disciplines. In social insects, division of labor (DOL) is an emergent, colony-level trait thought to depend on colony composition. Theoretically, behavioral response threshold models have most commonly been employed to investigate the impact of heterogeneity on DOL. However, empirical studies that systematically test their predictions are lacking because they require control over colony composition and the ability to monitor individual behavior in groups, both of which are challenging. Here, we employ automated behavioral tracking in 120 colonies of the clonal raider ant with unparalleled control over genetic, morphological, and demographic composition. We find that each of these sources of variation in colony composition generates a distinct pattern of behavioral organization, ranging from the amplification to the dampening of inherent behavioral differences in heterogeneous colonies. Furthermore, larvae modulate interactions between adults, exacerbating the apparent complexity. Models based on threshold variation alone only partially recapitulate these empirical patterns. However, by incorporating the potential for variability in task efficiency among adults and task demand among larvae, we account for all the observed phenomena. Our findings highlight the significance of previously overlooked parameters pertaining to both larvae and workers, allow the formulation of theoretical predictions for increasing colony complexity, and suggest new avenues of empirical study.

Harmonic radar tracking reveals that honeybee drones navigate between multiple aerial leks

Article

Full-text available

May 2021

Male honeybees (drones) are thought to congregate in large numbers in particular “drone congregation areas” to mate. We used harmonic radar to record the flight paths of individual drones and found that drones favoured certain locations within the landscape which were stable over two years. Surprisingly, drones often visit multiple potential lekking sites within a single flight and take shared flight paths between them. Flights between such sites are relatively straight and begin as early as the drone’s second flight, indicating familiarity with the sites acquired during initial learning flights. Arriving at congregation areas, drones display convoluted, looping flight patterns. We found a correlation between a drone’s distance from the centre of each area and its acceleration toward the centre, a signature of collective behaviour leading to congregation in these areas. Our study reveals the behaviour of individual drones as they navigate between and within multiple aerial leks.

Social networks predict the life and death of honey bees

Article

Full-text available

Feb 2021

In complex societies, individuals’ roles are reflected by interactions with other conspecifics. Honey bees (Apis mellifera) generally change tasks as they age, but developmental trajectories of individuals can vary drastically due to physiological and environmental factors. We introduce a succinct descriptor of an individual’s social network that can be obtained without interfering with the colony. This ‘network age’ accurately predicts task allocation, survival, activity patterns, and future behavior. We analyze developmental trajectories of multiple cohorts of individuals in a natural setting and identify distinct developmental pathways and critical life changes. Our findings suggest a high stability in task allocation on an individual level. We show that our method is versatile and can extract different properties from social networks, opening up a broad range of future studies. Our approach highlights the relationship of social interactions and individual traits, and provides a scalable technique for understanding how complex social systems function.

The neuroethology of olfactory sex communication in the honeybee Apis mellifera L

Article

Full-text available

Jan 2021
CELL TISSUE RES

The honeybee Apis mellifera L. is a crucial pollinator as well as a prominent scientific model organism, in particular for the neurobiological study of olfactory perception, learning, and memory. A wealth of information is indeed available about how the worker bee brain detects, processes, and learns about odorants. Comparatively, olfaction in males (the drones) and queens has received less attention, although they engage in a fascinating mating behavior that strongly relies on olfaction. Here, we present our current understanding of the molecules, cells, and circuits underlying bees’ sexual communication. Mating in honeybees takes place at so-called drone congregation areas and places high in the air where thousands of drones gather and mate in dozens with virgin queens. One major queen-produced olfactory signal—9-ODA, the major component of the queen pheromone—has been known for decades to attract the drones. Since then, some of the neural pathways responsible for the processing of this pheromone have been unraveled. However, olfactory receptor expression as well as brain neuroanatomical data point to the existence of three additional major pathways in the drone brain, hinting at the existence of 4 major odorant cues involved in honeybee mating. We discuss current evidence about additional not only queen- but also drone-produced pheromonal signals possibly involved in bees’ sexual behavior. We also examine data revealing recent evolutionary changes in drone’s olfactory system in the Apis genus. Lastly, we present promising research avenues for progressing in our understanding of the neural basis of bees mating behavior.

Functional Heterogeneity in Superorganisms: Emerging Trends and Concepts

Article

Full-text available

Nov 2020
ANN ENTOMOL SOC AM

Social insects are biological benchmarks of self-organization and decentralized control. Their integrated yet accessible nature makes them ideal models for the investigation of complex social network interactions, and the mechanisms that shape emergent group capabilities. Increasingly, interindividual heterogeneity, and the functional role that it may play, is seen as an important facet of colonies’ social architecture. Insect superorganisms present powerful model systems for the elucidation of conserved trends in biology, through the strong and consistent analogies that they display with multicellular organisms. As such, research relating to the benefits and constraints of heterogeneity in behavior, morphology, phenotypic plasticity, and colony genotype provides insight into the underpinnings of emergent collective phenomena, with rich potential for future exploration. Here, we review recent advances and trends in the understanding of functional heterogeneity within social insects. We highlight the scope for fundamental advances in biological knowledge, and the opportunity for emerging concepts to be verified and expanded upon, with the aid of bioinspired engineering in swarm robotics, and computational task allocation.

Asian Honey Bees: Biology, Conservation, and Human Interactions

Book

Dec 2006

The Major Transitions in Evolution

Book

Oct 1997

Over the history of life there have been several major changes in the way genetic information is organized and transmitted from one generation to the next. These transitions include the origin of life itself, the first eukaryotic cells, reproduction by sexual means, the appearance of multicellular plants and animals, the emergence of cooperation and of animal societies, and the unique language ability of humans. This ambitious book provides the first unified discussion of the full range of these transitions. The authors highlight the similarities between different transitions--between the union of replicating molecules to form chromosomes and of cells to form multicellular organisms, for example--and show how understanding one transition sheds light on others. They trace a common theme throughout the history of evolution: after a major transition some entities lose the ability to replicate independently, becoming able to reproduce only as part of a larger whole. The authors investigate this pattern and why selection between entities at a lower level does not disrupt selection at more complex levels. Their explanation encompasses a compelling theory of the evolution of cooperation at all levels of complexity. Engagingly written and filled with numerous illustrations, this book can be read with enjoyment by anyone with an undergraduate training in biology. It is ideal for advanced discussion groups on evolution and includes accessible discussions of a wide range of topics, from molecular biology and linguistics to insect societies.

The Dancing Bees: An Account of the Life and Senses of the Honey Bee

Book