Assessing the various anthropogenic pressures imposed on honey bees requires characterizing the patterns and drivers of natural mortality. Using automated life-long individual monitoring devices, we monitored worker bees in different geographical, seasonal and colony contexts creating a broad range of hive conditions. We measured their life-history traits and notably assessed whether lifespan is influenced by pre-foraging flight experience. Our results show that the age at the first flight and onset of foraging are critical factors that determine, to a large extent, lifespan. Most importantly, our results indicate that a large proportion (40%) of the bees die during pre-foraging stage, and for those surviving, the elapsed time and flight experience between the first flight and the onset of foraging is of paramount importance to maximize the number of days spent foraging. Once in the foraging stage, individuals experience a constant mortality risk of 9% and 36% per hour of foraging and per foraging day, respectively. In conclusion, the pre-foraging stage during which bees perform orientation flights is a critical driver of bee lifespan. We believe these data on the natural mortality risks in honey bee workers will help assess the impact of anthropogenic pressures on bees.

To date, two main techniques are available to monitor the in-and-out activity of honey bees at the beehive entrance, the Radio-Frequency Identification device (RFID) (49) and the optical counter based on image identification of barcode tags (50, 51). With such devices, the time-activity budgets of flights (duration and number of trips per day) are similarly recorded at the individual level. We equipped three colonies with RFID in 2011 in the Long Term Social-Ecological Research “Zone Atelier Plaine & Val de Sèvre” (LTSER ZA-PVS) (52) in central western France (46°23’N-0°41’W, site A, Figure 1), and one colony with image-based optical counters in 2018 at the National Institute for Agriculture, Food and Environment of Avignon in southeastern France (43°54’N-4°-52’E, site B, Figure 1). The RFID device consists of two adjacent rows of five contiguous RFID readers (iID2000, 2k6 HEAD; Microsensys GmbH, Erfurt, Germany) placed at the entrance of the hives (Figure 1). The image-based optical counter comprises a camera that monitors the hive entrance and image analysis software that detects and registers the barcode (50, 51). These two geographical areas vary in climate context (Oceanic for site A and Mediterranean for site B), giving different colony dynamics (e.g. different timing of peak brood production).

Beyond this inter-regional variability, we further introduced intra-regional variations in bee life-history traits (Figure 1). To do so, cohorts of bees were tagged each month from April to September in both sites to add seasonal variation. In addition, we simulated a third factor of variation by manipulating the environment of tagged bees (Figure 1). In site A, all tagged bees came from a single source colony (A. mellifera mellifera × caucasica strain) but were introduced into three monitored beehives (common 10-frame Dadant-Blatt model with sister queen of A. mellifera mellifera × caucasica strain) placed  about 30 km from the source location. Moreover, the three monitored colonies were separated from each other by about 15km to provide independent foraging landscapes of the tagged bees. In site B, tagged bees that were introduced each month in a single monitored colony came from three source colonies (A. mellifera mellifera × ligustica strain for the source and monitored colonies). Hence bees could be identified based on their colony of origin. These two different scenarios allowed the creation of variability in environmental and colony contexts with potential effects on bee life history.

To monitor individuals throughout their entire adult life, we collected newly emerged bees. Bees were selected from brood frames containing late-stage pupae, all adult bees were removed from this frame and placed in an incubator at 34°C and 50–70% humidity. Newly emerged bees were then collected after 2-10 hours of incubation and tagged. In site A, tags consisted of RFID microchips of 1.0 mm × 1.6 mm × 0.5 mm (mic3® - TAG 64 bit RO, iID2000, 13.56 MHz system, Microsensys GmbH, Erfurt, Germany). RFID tags weighed approx. 3 mg, i.e. 3 % of an adult honey bee’s body mass. This weight is considered low enough to not interfere with the individual life and tasks (49). In site B, tags consisted of data-matrix barcodes (3mm diameter) printed on laminated paper. Both tag types were glued on the bee thorax using biocompatible dental cement (TempoSIL2®) in site A and Sader® glue in site B. Because bee weight might influence flying capabilities (46), in site B, all newly emerged bees were weighed individually before tagging in order to assess the potential influence of body weight at emergence on bee lifespan. Indeed, bee weight at emergence might vary according to their nutritional state during larval development (53).

Tagged bees (less than 24h old) were then introduced into colonies equipped with monitoring systems and monitored until death in both sites. All colonies were adequately treated against the mite Varroa destructor each fall and no visible disease symptoms were observed.

Notes are provided in the R code.