iManagement

Bee behavior within the hive: insights from a long-term video analysis

The individual behavior of the honey bee is embedded in the overall management of the colony as a superorganism. Activity deep within the brood nest is now revealed thanks to the work of a research team that has developed a technology for digitally recording striking video footage. The intimacy of the colony is exposed down to the bottom of the comb cells: egg laying, larval hatching, care of open brood, and food storage appear in short videos, making it possible to visualize behaviors that had previously gone unnoticed.

Résumé / Abstract

The combined behaviors of individuals within insect societies determine the survival and development of the colony. In the western honey bee (Apis mellifera), individual behaviors include nest construction, foraging, food storage and maturation, brood nursing, temperature regulation, hygiene, and defense. However, the various behaviors inside the colony—especially inside cells—are hidden from view and, until recently, were mainly described through text and line drawings lacking the dynamics of moving images. In this study, we provide a comprehensive online source of video material offering a view of honey bee behavior inside comb cells, thereby providing a new mode of observation for the scientific community and the general public. We analyzed long-term video recordings of cells cut longitudinally, which allowed us to look laterally into cells within the middle of a colony. Our qualitative study provides insights into worker behaviors, including the use of wax scales and existing nesting material to remodel comb, storing pollen and nectar in cells, brood care and thermoregulation, and hygiene practices such as cannibalism, grooming, and surface cleaning. We reveal unique processes that have not been published previously, such as the rare mouth-to-mouth feeding of larvae by nurse bees as well as thermoregulation within cells containing developing brood. Using our unique video method, we are able to present the processes of a fully functional colony of social insects in classrooms and homes, thereby facilitating ecological awareness in modern times. We provide new details and imagery that will help scientists test their hypotheses about social behaviors. In addition, we encourage non-commercial use of our material to educate beekeepers, the media, and the public and, in return, to draw attention to the general decline in insect biomass and diversity.

 

Introduction

The survival, progress, and homeostasis of a honey bee colony depend on the coordination of advantageous decisions by individuals. The complex social organization of honey bees and other hymenopteran insects has been the subject of many studies (e.g. [ 1 , 2 ]). This research, whose history spans several centuries, has addressed division of labor such as comb construction, foraging, food storage and maturation, brood nursing, temperature regulation, and hygiene. However, because most of these behaviors are hidden from view, no educational video material has existed to date.

Darwin took important steps to educate the public about bee behavior, describing the remarkable comb-building activities of honey bees in his writings [ 3 ]. Similarly, in the early 19th century, Huber observed nest construction through a glass-walled observation hive that opened like a shelf, allowing him to view the colony [ 4 ]. Huber’s observation hive was based on one conceptualized by de Réaumur, who, with his new construction, investigated honey bee behavior through a glass surface [ 5 ]. These efforts to see inside the hive established one of the fundamental aims of video observations inside the hive.

Until the 21st century, educational material available on bee behavior was almost exclusively limited to texts and illustrations. For example, studies using observation hives were cited in works describing the following processes: the production and use of wax scales for comb construction [ 6 , 7 ], the regression of foragers to nurses despite advanced age [ 8 , 9 ], and the various dances bees use to communicate [ 10 – 13 ]. Now that online streaming platforms and digital recording technology allow broad dissemination of educational videos, bee behavior should be made available in this form.

While various behaviors can be observed outside comb cells, those within cells—such as brood care—are more difficult to observe because the view is blocked by bees covering their respective cells. For example, in the early 20th century, larval development could only be examined by extracting larvae from cells [ 14 ]. A solution to this problem was first proposed by Martin Lindauer [ 15 ]. By rotating comb strips by 90°, regulating temperature loss with a double layer of glass, and leaving only a small space for comb construction, he induced bees to rear brood in cells with translucent cell walls. After observing these cells, Lindauer described the nursing process in writing, whereas 35 years later an analog video recording device was used for the first time for this purpose [ 16 ]. However, this video material never became accessible to the public, as the internet was still in its infancy.

In parallel with the development of new video techniques providing a lateral view into cells, scientists discovered a previously unobservable behavior: active thermoregulation by honey bees inside cells. Infrared cameras revealed that workers showing very little movement were not resting but heating the comb from inside cells [ 17 ]. This discovery complemented other important aspects of thermoregulation that had been extensively described in the past [ 18 - 23 ].

In a recently published study [ 24 ], we combined our method of looking laterally into cells with long-term digital recordings. We continuously recorded brood areas of our observation hives, resulting in detailed views of a wide range of honey bee behaviors and offspring development. While that study focused on the impact of neonicotinoids on nursing behavior, here we present quantitative and qualitative analyses of social behaviors observed in these long-term recordings. These analyses include quantifying brood-cell visits and many video sequences of worker behaviors such as the creation and use of wax scales, deployment and uptake of pollen and nectar, brood care and inspection, thermoregulation, capping, cannibalism, grooming, and surface cleaning. Our footage also shows embryonic hatching and larval cocooning within the colony. Furthermore, we reveal in detail several previously undescribed behaviors, including comb remodeling, exceptional mouth-to-mouth feeding between a nurse and a larva, and pollen trapping by foragers. In addition, we further elucidate the method of water evaporation in brood cells. For the first time, we provide online, publicly accessible recordings of each of the above-mentioned behaviors for educational purposes.

 

 

Video: How bees care for brood

 

Materials and methods

Hives and recording setup

We used small observation hives, each with a population of approximately 3,000 individuals (300 g of Apis mellifera carnica) and a queen. Bees and sister queens were taken from the hives and the local queen breeding program of the Institut für Bienenkunde in Oberursel, respectively. In the designated brood area, combs were rotated by 90°, allowing a view into longitudinally cut cells through 4 mm anti-reflective glass and continuous recording of adult behaviors and offspring development inside. We either recorded the complete brood area using 5.3 MP cameras with 12.5 mm focal length lenses (camera PL-D725MU-T, PixeLINK, Ottawa with lens LM12HC, Kowa Optical Products Co., Ltd., Tokyo; or camera acA2500-60uc, Basler AG, Ahrensburg with lens TS1214-MP, Basler AG, Ahrensburg) or recorded in macro format using a 5 MP camera with a 25 mm focal length lens (camera acA2440-75um, Basler AG, Ahrensburg with lens TS2514-MP, Basler AG, Ahrensburg). While approximately 420 cells were observable with relatively low spatial and temporal resolution (1 to 3 frames per second) in the full brood-area recordings, the macro recordings covered a section of approximately 8 cells with high spatial and temporal resolution (25 to 30 frames per second). For illumination, we used domes emitting red light beyond the honey bees’ color vision range (λpic = 660 nm) with a diameter of 36 cm for overview recordings and 20 cm for macro recordings. StreamPix (versions 6.3.0.155 and 7.4, NorPix Inc., Montreal) was the recording software used.

 

The breeding area of an observation hive was illuminated by a dome emitting red light beyond the color vision of honey bees. The dome was a 20 cm diameter metal serving bowl painted with matte white varnish on the inside and had a large hole drilled at the top for the camera lens.

 https://doi.org/10.1371/journal.pone.0247323.g001

 

Fig 1. Macro video recording setup.

 

Quantitative analysis of cell visits during worker development

We used data and basic principles of the method described in [ 24 ] to examine the duration and depth of worker visits into a brood cell in the present study. We projected videos (2D + time) into images (1D + time), concatenating the brightness of the central pixels of cells from bottom to entrance over time (see https://www.nature.com/articles/s41598-020 -65425-y/chiffres/1 ). Bees were darker than the surrounding cell wax, and using a gray threshold we were able to detect events. For each event, the algorithm described in [ 24 ] determined the total duration. Temporal resolution was one second in these experiments (recordings of all visible cells in the brood area).

 

Results and discussion

 

1. Egg movements and larval hatching

After oviposition ( Video S1 ), the egg remains immobile until larval hatching.

 

Video S1. Honey bee worker development: oviposition.

 

As workers move as deeply as possible into cells, eggs may be pushed toward the base of the cell ( Video S2 ).

 

Video S2. Bee thermoregulation: cell occupation.

 

Workers can move into cells to preserve (clustering) or create (direct incubation) heat within the comb, and in the process the worker and its antennae remain immobile. This observation is consistent with earlier suggestions that “tilting” is not part of the normal process of embryonic development [ 25 ] (cf. no tilting in video S3 in [ 24 ]]). 

 

Video S4. Honey bee worker development: larval hatching.

 

During hatching, egg membranes are fully dissolved [ 25 ]. The first feeding occurs 95.2 ± 11.3 (mean ± SEM ; n = 86) minutes after larval hatching.

 

2. Inspection, feeding, and cocooning of larvae

 

 

Inspections involve acquiring and processing sensory information to determine cell contents, location, condition and age of brood, etc. The main characteristic of inspections is frequent antenna movement. During warming or resting behaviors, which can be distinguished by the frequency of abdominal pumping movements [ 17 ], no antennal movement of the worker is present ( Fig 2 , Video S2 ). Inspections not followed by other behaviors occur either for very short durations, during which the worker barely enters the cell, or for relatively long durations, which is more common in cells with very young larvae ( Video S3 ).  

Fig 2. Differences in head alignment for different tasks upon entering a cell.

All drawings in this article are by Nastasya Buling.

 https://doi.org/10.1371/journal.pone.0247323.g002

 

In our quantitative analysis of cell visits in a cell containing developing larvae, the mean number of short events (< 10 seconds) was tenfold higher than events of longer duration ( Fig 3B ). Across the six days of larval development (from first to last feeding), we detected 13,972 ± 617 events in the cell (mean ± SEM, n = 52 in 3 colonies). The number of events increased from 2,129 ± 186 on the first day of larval development to 2,497 ± 206 on the second day ( Fig 3A). Similar numbers were counted on the third (2,461 ± 154) and fourth (2,436 ± 115) day. We observed the most events on the fifth day (3,065 ± 143) and the fewest on the sixth day (1,383 ± 154). This pronounced decrease in events on the last day of larval development was caused by the increasing likelihood that workers obstructed the cell. In contrast, mean visit duration was highest on the first (17.0 ± 3.3 seconds) and lowest on the last day of larval development (5.3 ± 0.3 seconds).

 

Larval development was divided into six larval development days. A. While the mean number of visits increased from the first to the fifth development day, mean visit duration decreased by two thirds over this period. B. Short-duration visits (< 10 seconds) were tenfold higher than those combi


See also:

Author
Paul Siefert , Nastasya Buling, Bernd Grünewald
Back to overview