Experiment 1
To study the acceptance and tag retention rates of honey bees under different introduction conditions, we set up three two-frame observation hives with (sim 1500) adult bees and a queen. Observation hives were set up in a shed, with an entrance tube that connected to the outside 4 ft above the ground so bees could freely forage in the surrounding fields (Fig. 1A). We put emerging brood from healthy source colonies in an incubator ((33.5^{circ }) C, (ge 55)% RH) and tagged individuals that emerged overnight. Each hive had two vent holes (1” Dia.) through which we could introduce bees (Fig. 1A).
(A) Observation hive with introduction holes (red), through which bees were introduced via funnel or introduction cage. (B) Plastic tags, silicon tags, and sucrose spray. (C) Photograph of a tagged bee foraging (Photo by Greg Yauney).
In our initial experiment, there were seven treatment groups with 20 bees each per colony (n = 420 bees total). The seven treatments were: Control (C), No Sucrose (NS), Plastic (P), Wood Glue (WG), Not Incubated (NI), No Cage (NC), and Day (D). The Control treatment was designed to be a positive control, where we applied all the techniques we thought might increase acceptance of bees into a colony and tag retention rates. All other treatments had a single difference in the tag, tagging process, or introduction process to distinguish it from the control group, as detailed in Table 1. We glued a tag to the thorax of each bee (Fig. 1B) and marked the abdomen with a paint pen (Posca) to distinguish among treatment groups. In order to glue tags on, we picked each bee up, placed a small amount of glue on the thorax, and placed a tag on top of the glue with a pair of forceps (see Video 1 which details the tagging process). All bees except those in the Plastic group were tagged with 1.7 mm(^2) silicon tags (3.4 mm area). Silicon was chosen because it is a material representative of ASICs, which you would expect in a custom chip designed to track bee foraging flights. Plastic tags were 3 mm Dia. plastic discs (7.07 mm area), which are the commercially available bee tags commonly used in honey bee tracking and behavior experiments (Betterbee). All tags were glued on with shellac glue, the glue that comes with commerical honey bee marking kits (Betterbee), except for in the wood glue group, where they were glued on with wood glue (Titebond III). Next, bees were placed in a container with a bit of honey and stored until they were ready to be introduced. All bees except those in the Not Incubated group were placed in the incubator ((33.5^{circ }) C, (ge 55)% RH). The Not Incubated group was stored in a room environment, with variation between 21–27(^{circ }) C and 35–42% RH until introduction. Bees in the Day treatment spent 5 h in the incubator and then were sprayed with a light sucrose syrup (1 sucrose: 1 water (v/v)) and introduced at 4pm while the hives were still actively foraging. The rest of the bees spent between 5 and 8 h in the incubator or room environment before being introduced at 10:30 pm, after foraging had concluded. All except the No Sucrose group were sprayed with a light sucrose syrup before being introduced. The No Cage bees were rapidly introduced through one of the vent holes on the top of the hive using a funnel. The rest of the bees were placed in a cage together, which we connected to the introduction holes at the top of the colony, allowing them to move freely between the cage and the hive.
Beginning on day 2 (07/09/2020), we observed each hive in the morning on days 2-4 and 6-9 to see how many bees per group were present, hereinafter referred to as presence, and how many bees per group were present with tags, hereinafter referred to as success. We selected a random order in which to observe the three hives and a random order in which to observe the treatment groups for each hive. Each side of each hive had a grid drawn on it that divided it into nine squares. We scanned each side of each colony by eye for each treatment, starting with the lower left square of the grid on the first side, moving across the row, and then moving up to the next row, counting presence and success, using a tally counter when needed. We then moved rapidly to the other side and started at the top left of the grid, scanning row by row until we had observed each square in the grid. After an initial scan for each treatment, we placed the covers on the hives and shook for 10s to encourage bees to move around in the hive, and then waited for at least 15 minutes before a second observation. The maximum presence and success from the two daily observations were used for each treatment group and hive for analysis. Since we collected data by scanning each colony, we sometimes found more bees from a group in an observation hive than we had found in the same hive on previous day(s), even though more time had passed. Over the course of the experiment, our hives grew in size, and we believed we were seeing less tagged bees in part because they made up a smaller proportion of the hive population, and so decided to do a destructive sampling before the tagged bees reached foraging age. After dark on day 14 (7/21/2020), we made sure no tagged bees were dead on the bottom of the hives. We blocked the entrances, vacuumed all bees at the entrances into containers, and froze vacuumed bees and the three colonies, so that we could do a destructive sampling of all 3 colonies. This allowed us to get a final count of the presence and success for each of the seven treatment groups. We dissected each frozen colony, removing and inspecting each dead bee, and recorded the presence and success of each treatment group.
Experiments 2 and 3
We set up three two-frame observation hives in the same shed used for experiment 1 to conduct follow-up experiments in August 2020. The goal of experiment 2 was to compare Gorillaglue gel, an easily accessible Superglue (SG), to Titebond III, a readily accessible Wood Glue (WG2) used in experiment 1. We placed frames of capped brood in an incubator overnight to produce one day old nurse bees. We picked up each bee, placed a small dot of either superglue or wood glue on the thorax, and then placed 1.7 mm(^2) silicon tags on top of the glue. Bees were stored in the incubator with honey for 5–6 h until after dark. Then, we sprayed the bees with a light sucrose syrup and connected their cages to the vent holes at the top of the observation hives, allowing the bees to freely move between their cage and the hives. These details are summarized in Table 1.
Some honey bee tagging projects may benefit from tagging foragers as opposed to nurse bees, because nurse bees are the youngest workers and if you tag them you must wait for them to reach foraging age, during which time they may lose their tags. Specifically, tagging foragers as opposed to nurses will be advantageous when the tag price is extremely high or the project is very time constrained, and knowing the exact age of tagged bees is not important for the project goals. Since foragers are older workers that have already acquired the colony scent and learned to navigate the area surrounding their hive, the optimal methods for introducing nurses and foragers may differ. It is not easy to use bees from a source colony, because if they are within foraging range of their maternal colony, they will attempt to fly back home. The goal of experiment 3 was to apply a treatment that had high success with nurse bees (Experiment 1: WG) to foragers, and compare with releasing foragers near their colony and allowing them to return freely. We call these treatments Hive Introduced (HI) and Natural Release (NR), respectively. All foragers for this experiment were collected from the observation hives and were introduced back to the same observation hive after tagging, either through the vent holes at the top of the hive or by releasing the bees near the entrance of the hive. We collected foragers from each colony entrance into a cage with an insect vacuum (Hand-Held DC Vac/Aspirator, Bioquip), specifically aspirating bees that were arriving from foraging trips or had nectar loads, and placed them in the fridge to anesthetize them. We then selected those with intact wings, placed a dot of wood glue on their thoraxes, and placed silicon tags on top of the glue. Both treatment groups were stored in the incubator ((33.5^{circ }) C, (ge 55)% RH) and given honey to feed on. After 2 h in the incubator, the containers with NR bees were sprayed with a light sucrose syrup and placed on the ground 5 ft in front of their respective hive entrances and opened, allowing the bees to fly back to their hives unaided. At 10PM, when it was dark and foraging had concluded, the HI bees were sprayed with a light sucrose syrup. Their cages were then connected to the vent holes at the top of the observation hives, allowing them to freely move between their cage and the hives.
Experiments 2 and 3 were conducted in the same hives simultaneously, but were considered separate experiments because experiment 2 was conducted with nurses of known age and experiment 3 was conducted with foragers of unknown age. Nurses and foragers typically have an age difference and experience different levels of risk due to the behaviors they engage in, and so we analyzed these data separately in order to not confound our results. Beginning on day two (08/26/2020), we observed each hive on days 2–11 and 15–21 to determine introduction presence and success for experiment 2 and experiment 3. Forager observations (experiment 3) were always done early in the morning, before foraging activity commenced. As in experiment 1, we randomized observation order, scanned colonies for each treatment group before and after shaking, and used the maximum presence and success from the two observations for analysis. Since we collected data on multiple days by scanning each colony, we occasionally found more bees in a group than we had found on previous day(s), even though more time had passed.
Statistical methods
Statistical analyses were performed in R 4.0.520. To determine which preparation and introduction techniques were associated with the highest presence and success, we built generalized linear mixed-effects models (glmms)21 for the proportion of present and success bees to introduced bees respectively, with treatment and sampling day as fixed effects, and colony as a random effect. For experiment 1, treatment was a categorical variable, where the Control bees were the reference group. We assessed the significance of the full models using Wald likelihood ratio chi-square tests on each glmm (‘Anova’ function in the ‘car’ package with test set to ‘Chisq’)22. In all statistical tests, (alpha) was set to 0.05. The destructive data from experiment 1 were analyzed separately from hive observation data. We ran a correlation test to determine the relationship between hive observation data from the final observation day, day 9, and the destructive sampling on day 14 using the ggpubr package23.
Source: Ecology - nature.com
